AI Trends and Predictions 2025 From Industry InsidersAI Trends and Predictions 2025 From Industry Insiders

This time last year, AI — generative AI, specifically — was mostly hype. A lot has changed in one year. As one industry insider put it: "In 2023, organizations were exploring and experimenting, and in 2024, they were implementing AI at scale. Because of the widespread implementation, in 2025, we will see an emphasis on ROI." Another calls generative AI the most important tech trend of 2025.

Our 2025 tech predictions are in, complete with "anti-predictions" — highlighting trends widely expected to dominate the IT landscape but viewed differently by our experts. Not surprisingly, a number of our predictions revolve around artificial intelligence — including a bearish outlook about expanded AI adoption by businesses in 2025.

Now it's IT leaders' and industry insiders' turn to share what they are expecting from AI in 2025. Check out their predictions below:

AI Joins the Dev Team

2025 will be the year developer capacities are enhanced by the power of AI, as AI tools are officially integrated into the developer tech stack. Over the course of the next year, we'll see team structure and processes adapt to maximize collaboration between AI and developers, experimenting with AI-augmented workflows and increased automation of demanding responsibilities like on-call shifts to supercharge efficiency and velocity. — Matt Makai, VP of Developer Relations & Experience, LaunchDarkly

Related:Top 10 Articles About AI of 2024

AI Will Transform Storage

In 2025, AI will continue to proliferate across all industries, driving new opportunities and challenges. The integration of AI into storage systems will be particularly transformative, with AI-powered solutions becoming increasingly common for optimizing performance, enhancing security and ensuring data reliability. This increase in AI workloads will lead to a surge in demand for high-performance storage solutions that can support these data-intensive applications, including large language models (LLMs), machine learning model training, and real-time data analytics. This will increase the requirements to data storage technologies in order to handle AI's specific needs for speed, scalability and efficiency. — Boyan Ivanov, CEO, StorPool Storage

The Rise of AI-Driven Sales Agents

In the next 12-18 months, we will see the rise of AI-driven agents in selling B2B goods in manufacturing, and B2B and B2C goods in retail, setting the foundation for other industries to experiment with them. These agents will understand the needs of the buyer, will be trained to interact with a seller's website using machine-to-machine protocols, and will have the ability to accurately source the right color/size/price at an unprecedented pace. Once these agents move from the idea stage to successful deployments to mainstream use, it will significantly change the way products are sold online, removing unnecessary human interactions and enabling people to focus on more impactful activities. — Jonathan Taylor, CTO, Zoovu

Related:AI Basics: A Quick Reference Guide for IT Professionals

Growing Convergence of AI, AppSec, and Open Source

We will see the continued intersection of AI, AppSec, and open source — from malicious actors targeting open source models, the communities and platforms that host them, and organizations looking to leverage AI to address code analysis and remediation. Increasingly, we will see widely used OSS AI libraries, projects, models, and more targeted as part of supply chain attacks on the OSS AI community. Commercial AI vendors are not immune either, as they are large consumers of OSS but often aren't transparent with customers and consumers regarding what OSS they use. — Chris Hughes, chief security advisor at Endor Labs

Related:AI Quiz 2024: Test Your AI Knowledge

AI Asset Management Challenges Emerge

Asset management challenges are coming to AI. As businesses start building out a catalogue of models, they will encounter challenges with size, portability, and discoverability. The industry will look for ways to get better compression with minimal reduction in accuracy from these assets so that they are more portable. There will be a need to effectively manage models, making them easy to find across organizations, and making them ever more interoperable. — Robert Elwell, VP of engineering, MacStadium

Ransomware and Digital Extortion (R&DE)

We expect R&DE incidents to continue at an elevated level in 2025, representing a significant threat to organizations of all sizes, industries, and geographies. 2024 was a record year for R&DE collectives with ZeroFox identifying an average of 388 incidents each month throughout 2024, compared to an average of 337 per month in 2023. Organizations that make up the manufacturing industry are likely to face the biggest threat from R&DE actors throughout 2025, with those within the retail, construction, healthcare, and technology sectors also prone to high levels of targeting. The greatest threat in early 2025 will very likely emanate from RansomHub, an extortion collective that was first observed in early 2024 and went on to become the most prominent R&DE outfit of the year. 2025 is likely to see an increasing number of new threat collectives, which continue to diversify the R&DE threat landscape. New collectives will also continue to develop and test new TTPs, such as increased emphasis on data extraction over traditional encryption methods, and to opt for double or triple extortion tactics in a bid to increase the chance of successful ransom demands. — Adam Darrah, VP of intelligence, ZeroFox

Geopolitical & Cyber Convergence

The cyber threat landscape in 2025 is expected to be heavily influenced by geopolitical developments, continuing the trend of increasing convergence between cyber and geopolitical spheres. Throughout 2024, geopolitical events directly impacted the motivations, capabilities, and intentions of cyber threat actors, including nation-state cyber capabilities, financially motivated DDW actors, ideologically motivated hacktivist collectives, and politically motivated activist groups. The dynamic and unpredictable geopolitical environment is expected to further influence cyber threat activities in 2025. Past and ongoing geopolitical events, such as the Russia-Ukraine war and the Israel-Hamas conflict, have facilitated elevated cyber threat activity. In 2025, we anticipate continued politically motivated cyber threats, including social engineering, data breaches, DDoS attacks, and malicious payload deployment, such as R&DE and spyware. Cybercriminal collectives are likely to align with geopolitical disputes, contributing to the complexity of the threat landscape. The EU's investment in high-tech fields and the geopolitical tensions between China, the US, and the EU are likely to intensify cyber threats, with state-backed actors targeting critical infrastructure and technology sectors. Russia and Iran are expected to use hybrid tactics, including cyber warfare, to advance their geopolitical agendas, further shaping the cyber threat landscape in 2025. — Adam Darrah, VP of intelligence, ZeroFox

Initial Access Brokers (IABs)

In 2025, Initial Access Brokers (IABs) are expected to remain a significant threat to organizations globally. The market for illicit network access surged in 2024, with record levels of IAB sales identified across DDW marketplaces. We anticipate this thriving market will continue in 2025, with IABs targeting organizations of all sizes, industries, and geographies. IABs sell unauthorized access to corporate networks by marketing compromised credentials and network entry points, allowing buyers to quickly exploit compromised networks with minimal investment and risk. The average purchase price of IAB sales in 2024 was under USD 5,000, offering substantial returns for threat actors, including R&DE collectives. The value of compromised access varies based on factors like information criticality, privilege level, and exploitation potential within the supply chain. Illicit access sales will likely continue underpinning the threat from R&DE operators in 2025, with security teams needing to be vigilant of IABs targeting them directly and indirectly via upstream partners. IABs are expected to focus more on third-party providers, perceiving them as having weaker security postures. North America will likely remain the primary target, followed by Europe, with industries like manufacturing, professional services, technology, retail, and financial services being the most attractive targets. — Adam Darrah, VP of intelligence, ZeroFox

First Major AI-Generated Code Vulnerability

Development teams have eagerly embraced AI, particularly GenAI, to accelerate coding and drive efficiency. While the push for the "10x developer" is transforming software creation, the need for speed can sideline or shortcut traditional practices like code reviews, raising significant security concerns. In the coming year, overconfidence in AI's capabilities could lead to vulnerable or malicious code slipping into production. GenAI is powerful but fallible — it can be tricked with prompts and is prone to hallucinations. This risk is not hypothetical: 78% of security leaders believe AI-generated code will lead to a major security reckoning. The CrowdStrike outage illustrated how quickly unvetted code can escalate into a crisis. With AI-generated code on the rise, organizations must authenticate all code, applications, and workloads by verifying their identity.

Code signing will become an even greater cornerstone in 2025, ensuring code comes from trusted sources, remains unchanged, and is approved for use. Yet, challenges persist: 83% of security leaders report developers already use AI to generate code, and 57% say it's now common practice. Despite this, 72% feel pressured to allow AI to stay competitive, while 63% have considered banning it due to security risks. Balancing innovation with security will be critical moving forward. — Kevin Bocek, chief innovation officer, Venafi, a CyberArk company

Everyone's a Creator: The Democratization of Specialized Knowledge Work

In 2025, AI tools will revolutionize knowledge work by enabling individuals to tackle tasks once reserved for specialists, from coding to design and content creation. Much like personal computers empowered workers to handle spreadsheets and documents independently rather than relying on centralized admin staff, AI will push creativity and productivity to the edge, placing advanced capabilities in the hands of individual contributors. This shift will not only accelerate workflows but also challenge traditional organizational structures as more people leverage AI to go solo or create in new ways. AI's role as a personal assistant and creative partner will reshape industries, making innovation more accessible than ever before. — Rob Brazier, VP of Product, Apollo GraphQL

AI-Driven APIs: A Wild Frontier

In 2025, the relationship between AI and APIs will enter uncharted territory, reshaping how systems are built and interact. AI will increasingly guide developers in crafting and consuming APIs, introducing new patterns and unpredictable usage scenarios. This shift will demand advanced observability tools to monitor and adapt to evolving behaviors, ensuring systems remain secure and efficient. As AI dynamically composes user experiences in real time, APIs will need to be more robust, resilient, and flexible than ever before. Businesses must embrace this wild frontier with innovation and foresight, as the synergy between AI and APIs transforms digital ecosystems in ways we're only beginning to understand. — Rob Brazier, VP of Product, Apollo GraphQL

AI and APIs: The Backbone of Intelligent Innovation

In 2025, the fusion of AI and APIs will redefine how businesses build and run intelligent systems. APIs will evolve from simple connectors to dynamic engines for innovation, driving experimentation and production at unprecedented scales. As AI applications proliferate, organizations will demand APIs that not only handle the chaos of rapid prototyping but also balance speed with robust security and cost efficiency in production environments. Granular access controls, real-time performance monitoring, and optimized compute environments will become non-negotiable for businesses navigating this new era. APIs will act as the trusted gatekeepers of sensitive data, ensuring that AI-driven systems are both powerful, smart, and secure. This synergy between AI and APIs will empower developers to build smarter, faster, and more resilient applications, setting a new standard for innovation across industries. — Subrata Chakrabarti, VP of Product Marketing at Apollo GraphQL

Smarter AI for Specialized Needs

In 2025, the future of AI will shift toward smaller, domain-specific systems designed to excel in targeted applications. These compact, context-rich models will redefine industries by offering unparalleled efficiency and precision. Rather than relying on broad, generalized AI, businesses will start to adopt solutions tailored to their unique needs — healthcare organizations will use AI for diagnostics, while financial institutions enhance fraud detection. By embedding deep and specialized knowledge directly into models, companies will deliver real-time insights and reduce resource demands. AI will take a step forward towards decision-making, serving as a critical assistant rather than a complete solution. This evolution will make AI more practical, accessible, and impactful, transforming specialized knowledge from an advantage into a necessity. — Subrata Chakrabarti, VP of Product Marketing at Apollo GraphQL

Ethics in AI Will Take a Step Forward in 2025

In 2025, geopolitical turbulence will continue, and misinformation is likely to abound. It's unlikely that new data privacy and AI policies will be passed and enforced in 2025, so customers will expect businesses to take responsibility for ethics in AI. As companies incorporate AI into their products, they have a responsibility to protect what and how the AI uses customer data, especially as it relates to sensitive data. Businesses must invest in ethical AI development, with an emphasis on transparency because AI adoption will directly correlate to the amount of trust the customers have in it. — Stephen Manley, CTO, Druva

2025 Will See the First Data Breach of an AI Model

Pundits have frequently warned about the data risks in AI models. If the training data is compromised, entire systems can be exploited. While it is difficult to attack the large language models (LLMs) used in tools like ChatGPT, the rise of lower-cost, more targeted small language models (SLM) make them a target. The impact of a corrupt SLM in 2025 will be massive because consumers won't make a distinction between LLMs and SLMs. The breach will spur the development of new regulations and guardrails to protect customers. — Stephen Manley, CTO, Druva

Synthetic Data Used More in AI Training to Safeguard Sensitive Customer Data, Creating New Risks

For AI to produce good results, it needs to be trained on good data and rigorously tested with prompt engineering. The business temptation is to use customer data to train AI models — but that causes a myriad of problems to crop up, such as data compliance breaches, higher impact of cyber risk, and higher likelihood of data leakage. To effectively combat these challenges, businesses will turn to synthetic data, or training data that AI models generate, to maintain safety best practices during the training process. This, however, will create new risks, since the synthetic data can create a feedback loop that will exacerbate any bias in the data. Therefore, companies will need to invest in transparency and increase the rigor in reviewing their AI-generated output. — Stephen Manley, CTO, Druva

2025 Is the Year of (Missing) ROI on GenAI Investments

The trough of disillusionment looms for GenAI, and the request for ROI will quicken the industry's descent into said trough. Every business is striving to understand the impact of GenAI, and savvy business leaders are already asking questions around accuracy, efficiency, and outcome to validate the IT spend allocated to it. Unless it's incorporated into a purpose-built tool from the ground up, GenAI won't drive significant measurable efficiency and many will feel let down by its initial promises. — Stephen Manley, CTO, Druva

Security Leaders Will Embrace AI Experimentation

2024 shocked many of us with AI technologies' sophistication and rapid advancement. The year also highlighted that we don't quite know how to incorporate such tools into work and which vendors can help us along the way. Organizations in 2025 will continue to experiment with AI to understand where it offers value. And we'll also see many startups experiment with business models and tech approaches. Security and IT leaders should be ready to help evaluate and onboard a diverse set of immature AI products. We'll need to comprehend a range of AI technologies and understand the expectations of diverse internal stakeholders so we can contribute toward making informed risk vs. reward decisions. — Lenny Zeltser, SANS Institute Fellow and CISO at Axonius

AI in Security: Balancing Human Expertise and Automation for Optimal Outcomes

AI-related advancements will continue to fuel discussions regarding the role of humans vs. automation in the workforce. Security teams will see more opportunities to use AI and non-AI technologies to automate tasks across many domains, including GRC, security operations, and product security. Security leaders will need to be strategic about deciding which tasks to leave for humans and which to automate. Given how rapidly the technology is changing, we should be ready to experiment and determine how to measure project outcomes to decide which approaches work best. — Lenny Zeltser, SANS Institute Fellow and CISO at Axonius

Multi-agent Neurosymbolic AI Will Advance Machine-to-Machine Collaboration

The first wave of multi-agent neurosymbolic AI applications that perform machine-to-machine collaboration will emerge in 2025. Agents across diverse systems — such as autonomous vehicles, robotics, and enterprise decision support platforms — will exchange and interpret complex symbolic representations of their surroundings in real time. These agents will work together to negotiate solutions, adapt to new situations, and coordinate actions based on both learned experiences and structured knowledge. This advancement will lead to a new wave of AI products capable of more intelligent teamwork and enhanced performance in complex environments, all while ensuring transparency and explainability in decision-making. — Dr. Jans Aasman, CEO, Franz

2025 Will Be the Year of the AI Agent

Instead of merely producing text or images, this new breed of AI application will be empowered to act. That might mean researching topics on the web, manipulating an application on a PC desktop, or any other task that can be performed via API. We're still a long way from general artificial intelligence, so these early agents will be quite specialized. We'll see the emergence of what might be called "agentic architectures" — focused use cases where AI can deliver immediate value. Likely examples include data modeling, master data management, analytics and data enrichment, where tasks are highly structured and prototypes have already shown promise. We'll see the first case studies in 2025, and then rapid uptake throughout the enterprise as lagging adopters see competitors gaining an edge. — Bob van Luijt, CEO, Weaviate

AI Moves Closer to the Edge  

In the year ahead, we anticipate AI at the edge will further enhance applications and improve efficiency with increasingly specialized edge-AI chips that can enable tasks with lower power consumption. AI techniques like TinyML and model quantization will continue to advance, allowing more sophisticated AI algorithms to run on resource-constrained devices. We expect more real-time speech recognition, computer vision, and predictive maintenance on small edge devices, along with more local data processing. Current edge applications mostly use pre-trained models, but a move toward real-time, on-device training and fine-tuning will become more common. This means edge devices could adapt and learn from local data over time, improving performance and personalization without relying on cloud retraining. — Rashmi Misra, chief AI officer, Analog Devices

Business Leaders Must Measure Value of AI Apps

Companies that rush into AI adoption without understanding their internal needs and bandwidth risk overwhelming their security and data teams with information that doesn't provide valuable insights. As AI continues to grow, businesses aiming for long-term ROI must shift their focus from simply integrating AI capabilities to addressing organizations' shortcomings and measuring value. To accomplish this in the coming year, business leaders should collaborate closely with internal teams to identify their processes, bottlenecks, and needs. By understanding these challenges, leaders can work strategically with their teams to determine the most effective AI applications and ensure their teams are prepared to manage them successfully. — Rishi Kaushal, CIO, Entrust

The 'AI Winter' Is Not Coming

We're currently experiencing one of the most sustained stretches of interest and investment in AI that we've ever seen. While traditionally we've seen this hype give way to "AI winters" where enthusiasm and funding taper off, this time around, there are strong indicators that this momentum will continue into the new year and beyond. 2025 will be a year where scaled production of AI will sustain the investment in AI for years to come. This is just the beginning. — Raj Pai, Vice President, Product Management, Cloud AI, Google Cloud

2025 Is the Year of the Platform

If 2024 was the year of the LLM, 2025 will be the year of the platform. There's no shortage of models on the market — plenty to address just about any use case. But there's no point for businesses in talking about models if you don't have a strong platform to support them. In 2025, technology leaders will shift their focus toward investing in platforms that have built-in security, grounding capabilities to reduce hallucinations, and can serve as a one-stop-shop to bring the potential of these models to life. — Raj Pai, Vice President, Product Management, Cloud AI, Google Cloud

AI Model Convergence Will Continue

If we look at the last 10 years of deep learning — now referred to as AI — we have been on a path of convergence. For example: In earlier days, we had separate models for different tasks like sentiment analysis, parts of speech tagging, and entity detection. But with models like BERT, a single model started performing all these tasks. Similarly: For translation, there were individual models for translating each language pair (i.e. English to Spanish, French to German, etc.). Now, a single model can translate across any pair of hundreds of languages. As we head into 2025, we'll see this convergence trend continue with things like screen understanding and reasoning, where a single model will have the power to do multiple tasks of varied nature, modalities, and across languages. With this strong technology march toward convergence, useful agentic behavior will natively start showing up across the different foundation models. — Saurabh Tiwary, VP, General Manager, Cloud AI, Google Cloud

Strengthening Cybersecurity Against AI-Generated Threats

With escalating threats from sophisticated phishing and ransomware attacks, focus needs to shift toward advanced data protection strategies, AI-driven threat detection and continuous employee training to mitigate ongoing risks. Businesses that proactively adopt these measures will not only comply with regulations but also build customer trust and loyalty. — James Tommey, Global Head of IT & Chief Security Officer, DISCO

Combating Fraudulent AI-Generated Content

In 2025, organizations will face unprecedented cybersecurity challenges due to the rise of fraudulent AI-generated content, which will become indistinguishable from human-created data. Leaders must think about how to implement robust authentication and verification protocols to safeguard against deepfakes and synthetic data breaches to ensure protection over the integrity of their workflows. — James Tommey, Global Head of IT & Chief Security Officer, DISCO

AI PCs Will Be a Hot Commodity

As the current PC refresh continues, AI PCs will be the primary choice in 2025 and beyond. We are only in the beginning stages of unlocking new workforce collaboration, security, productivity and even fulfillment through AI. For example, AI PCs now support real-time translation, extending global connection and collaboration seamlessly. As video communication cements itself as the norm, new eye-tracking capabilities allow participants to maintain perceived eye contact while focusing on facial cues and on-screen content, creating a more personal interaction. As business leaders evaluate their investments in new technology, AI PCs are a clear choice to begin equipping their workforce for the future with the high-performance computing required to scale and support increasingly data-driven and AI-powered work. — Dave McQuarrie, Chief Commercial Officer, HP Inc.

Breaking Down Data Silos Will Become a Central Focus for AI and Data Architects

In 2025, breaking down data silos will emerge as a critical architectural concern for data engineers and AI architects. The ability to aggregate and unify disparate data sets across organizations will be essential for driving advanced analytics, AI, and machine learning initiatives. As the volume and diversity of data sources continue to grow, overcoming these silos will be crucial for enabling the holistic insights and decision-making that modern AI systems demand. The focus will shift from the infrastructure toward seamless data integration across various platforms, teams, and geographies. The goal will be to create an ecosystem where data is easily accessible, shareable, and actionable across all domains. Expect to see new tools and frameworks aimed at simplifying data integration and fostering greater collaboration across traditionally siloed environments. — Molly Presley, SVP of Global Marketing, Hammerspace

GPU Demand Soars, but AI Adoption Has Companies Rethink Resource Allocation

As we enter 2025, the AI industry faces an unexpected situation: a huge demand for GPUs worldwide, yet many of these powerful chips aren't being fully used. While companies invested heavily in GPU-based infrastructure, many continue to struggle to apply these chips to AI workloads, instead redirecting them toward non-AI applications. The expected AI-driven boom remains slower than anticipated.

We will continue to see companies be more selective with GPU allocations, as companies focus on areas where the impact of AI in areas like data analytics and cloud computing enhancements — rather than emerging AI initiatives. Additionally, as developers become more resource-conscious, the focus on optimizing algorithms for available hardware, leveraging CPU-bound AI, and adopting hybrid approaches could become central trends. Ultimately, 2025 may be a year that companies will adapt to both the technical and logistical challenges of realizing AI's potential. — Molly Presley, SVP of Global Marketing, Hammerspace

Generative AI in 2025: A New Era of Innovation

As we move into the new year, I'm excited to see generative AI continue its rapid evolution, especially in areas where progress is already accelerating. Models focused on code and math (anything with well-defined reward signals) will become even more capable, pushing the boundaries of what we can automate and optimize. I expect open-weight models to reach a level of performance that makes them viable for a wide range of practical applications, making cutting-edge AI more accessible than ever before. Another area to watch is the growing role of AI-generated audio and video content. We will soon see this kind of content becoming a significant part of our everyday media consumption. I believe we're on the cusp of a major scientific breakthrough driven by AI, which will have profound implications for research and innovation. The pace of progress in generative AI is only going to accelerate, and I can't wait to see where it takes us next. — Percy Liang, co-founder, Together AI

Agentic AI Will Take Center Stage, Delivering on Personalization and Efficiency

In 2025, AI won't just be a tool; it will be a collaborator. Many AI-powered tools in use today are based on static rules or datasets. Agentic AI differs in that it can continuously learn from user inputs and integrate contextual information (think: account history, network environment, user behavior patterns and preferences), and make decisions with little to no human oversight. In other words, unlike today's approaches that require user prompts or predefined rules, agentic AI will operate proactively.

Imagine a customer service AI that predicts user needs before a query is made, or a network management AI that identifies potential issues and resolves them autonomously, ensuring uninterrupted service. 

These AI agents will not just interact with humans or devices directly, but will also be able to discover, learn, and collaborate with each other to form complex workflows and/or chains of operations to automate even advanced business functions. For instance, multiple AI agents could automate supply chain management by coordinating with each other to forecast demand, optimize inventories, coordinate deliveries, and even negotiate with suppliers. For businesses, this shift means a leap in efficiency and personalization. It also underscores the importance of governance and guardrails. In response to the rise of agentic AI, we will see organizations implementing mandatory ethical guidelines to ensure fairness and transparency in algorithmic decisions and protecting intellectual property. — Liz Centoni, Executive Vice President and Chief Customer Experience Officer, Cisco

AI Will Surface Tough Reality Checks for Companies

AI will continue to captivate businesses, promising unprecedented innovation and efficiency, and companies will continue to invest in AI-powered solutions. This is hardly a prediction. But as AI journeys progress, so too will the understanding that the path is fraught with hurdles. Despite billions of dollars invested into AI models and AI-powered solutions in 2024, new data from Cisco's AI Readiness Index shows that AI readiness has declined by one point globally over the past year — now only 13% of companies are ready to leverage AI-powered technologies to their full potential. 

In 2025 organizations will grapple with how best to secure the right level of compute power to meet AI workloads (today, only 21% of organizations say they have the necessary GPUs to meet current and future AI demands). Companies will need to lean on their strategic partners to identify and prioritize their AI use cases, upskill their teams, and modernize their infrastructure environments in a progressive, proportional way. IT teams will experience increasing pressure to optimize the management, hygiene, labelling, and organization of data, which is currently spread across multiple systems and locations. This mandate will apply to structured data typically associated with existing business processes, as well as unstructured data related to customer and user interactions. As teams work feverishly to prepare their environments for AI, boards and leadership teams will realize that significant gains from AI will happen in the long run and progressively — starting now and improving over time — especially in areas like opening new revenue streams and improving profitability. Many boards will find themselves readjusting expectations, timelines and priorities that were established mere months ago as companies reckon with the "messy middle" of AI implementation. Let's play the "long game." — Liz Centoni, Executive Vice President and Chief Customer Experience Officer, Cisco

Companies Will Need Help to Balance Sustainability and Growth in an AI-Powered Era

The environmental impact of AI is the elephant in a lot of rooms. AI requires high energy consumption levels that impact carbon emissions across the board. By 2025, the amount of energy used by data centers dedicated to AI is expected to match the amount consumed by a country the size of the Netherlands in one year. Indeed, in many of my AI conversations with customers, sustainability emerges as a core concern. In 2025, customers will increasingly seek out partners who can deploy technology while helping them meet their net-zero commitments and sustainability goals on their current timeline. Businesses that win will be those who prioritize energy-efficient products and circularity in business models. Interestingly, AI-powered technology could also play a crucial role in unlocking energy efficiencies. Businesses will see AI unlock a new era of "energy networking" that combines software-defined networking capabilities with an electric power system made up of direct current (DC) micro grids to deliver more visibility into emissions, and a platform for optimizing power usage, distribution, and storage. In 2025, AI will be both the "what" and the "how" in this space, bringing us vast capabilities and a continuous learning method for delivering them more sustainably. — Liz Centoni, Executive Vice President and Chief Customer Experience Officer, Cisco

AI Inches Closer to the Edge

2025 will be the year of real-time, multimodal AI. AI will enter the action with humans and machinery in entirely new ways — from bringing data from sensors, drones, robotics and machinery all together to take action. — Dan Wright, CEO and co-founder, Armada

Energy and Defense Reach an AI Tipping Point

In 2025, AI will hit its tipping point in energy, with edge computing bringing intelligence directly to the oil rig. Much like how railroads revolutionized the oil industry by unlocking new markets in the 19th century, cutting-edge computing infrastructure will transport AI to the farthest reaches of the edge in the 22nd century. 2025 will also mark a seismic shift in defense, as edge computing becomes indispensable in the era of autonomous warfare. It's the modern-day railroad that delivers AI to the frontlines, empowering the U.S. military to navigate the complexities of the battlefield with unprecedented speed and precision. — Dan Wright, CEO and co-founder, Armada

AI Will Enhance Customer Experience Management

This is the area where most companies are beginning their GenAI journey. They are trying out this new technology in a low-risk area to start. By offloading repetitive tasks requiring simple answers or informational lookups, companies seek to boost the customer experience with faster, more detailed answers through GenAI and RAG. — Adrienne Wilson, Director of Sales, Esker

Organizations Will Increasingly Use AI to Meet Sustainability Goals

In certain locations, companies are required to report on their sustainability results. By leveraging AI, automated suggestions can be made to choose a more sustainable option when procuring goods. Collecting and utilizing this data will help companies to meet these requirements. — Adrienne Wilson, Director of Sales, Esker

Business Leaders Develop More Mature AI Assessment Procedures

To enable business leaders to more effectively cope with the onslaught of "AI enabled" tools — and to minimize an oversight bottleneck — the industry will need to develop a set of foundational rubrics to guide in more timely assessments of AI technologies. As a result, I predict we will see a renewed focus on data classification labels, a better understanding of AI processing locations, and a demand for confidentiality assertions from vendors as private data traverses their infrastructure. As the industry transitions to an application-driven phase of AI, it is imperative that organizations be equipped to make thoughtful and timely decisions about how the technology can be used responsibly to drive business objectives. — Michael Covington, Vice President of Portfolio Strategy, Jamf

Schools Reinvigorate Efforts to Protect Students Online in the Wake of AI Proliferation

We'll see a strong push for more safety mechanisms to be installed on student devices, specifically when it comes to data protection, threat prevention, and privacy controls. Educational institutions will be encouraged (or perhaps required) to improve encryption protocols and access controls, use AI-powered threat detection to fight AI-powered attacks, use systems that provide real-time alerts, and step up their game when it comes to student data privacy. — Suraj Mohandas, Vice President, Strategy, Jamf

Enterprises Get a Reality Check on the Value of GenAI

Household name companies in cybersecurity to small new startups with 10 employees have quickly entered the GenAI market over the past year or two. It's a crowded space that can easily overwhelm even leaders of technology companies who are looking to select the right GenAI solution for their businesses. In 2025, while the hype cycle for GenAI will continue to evolve, we'll see the more effective solutions surface and more customers focusing on solutions that bring the most real value to their businesses. As with any "hot new tech" on the block, the buzz around this latest emerging technology will start to calm, and we'll start to see GenAI mature. We'll start to see what value these tools can provide for businesses, and which perform better than the others. It's going to be a year of cutting through the GenAI noise, and those who can break through that will be the companies that stick around for years to come. — Linh Lam, CIO, Jamf

To Open Source AI or Not? Navigating Innovation and Security Challenges

Open-source AI opens the door to unparalleled collaboration and innovation, but it also forces us to grapple with security, transparency, and trustworthiness questions. Organizations must weigh the benefits of openness against the potential risks of exposure. — Balaji Ganesan, co-founder and CEO, Privacera

Continued Proliferation of AI Use Cases

While AI isn't new, the momentum behind it is unprecedented. In 2025, we'll see a proliferation of AI use cases that redefine business processes. It's a transformative moment — some view it as a job risk, while others embrace it as an opportunity to innovate and thrive. — Sascha Giese, Global Tech Evangelist, Observability, SolarWinds

Balancing Innovation and Regulation: The Rise of Responsible AI Policies

Regulatory frameworks are stepping in to define the ethical, secure, and responsible path forward for AI and data usage. This is a wake-up call for organizations — compliance must transform from a checkbox exercise to a differentiating value proposition. Embracing these standards involves legal alignment and leading with purpose and integrity. — Balaji Ganesan, co-founder and CEO, Privacera

AI-Powered Predictive Maintenance and Risk Management to Dominate Building Systems

Managed services that monitor and optimize physical assets throughout their lifecycle will be table stakes. This includes critical functions like firmware updates, system health monitoring, and ensuring proper functionality. Predictive maintenance powered by AI will play a pivotal role in addressing vulnerabilities proactively, minimizing downtime and costs while bolstering security. The growing interconnectivity of building management systems brings new risks, including unvetted device access and limited visibility into system components. In 2025, facility managers need a layered risk management strategy that incorporates tiered system criticality, comprehensive remediation plans, and continuous auditing. — Greg Parker, Global Vice President, Security and Fire, Life Cycle Management, Johnson Controls

AI Will Bloom as Organizations Shift Focus to ROI and Efficiency

We've already seen all of the stages of the beginning "new technology" cycle with AI. In 2023, organizations were exploring and experimenting, and in 2024, they were implementing AI at scale. Because of the widespread implementation, in 2025, we will see an emphasis on ROI, and a focus on how AI can enable more efficient work. At this point, organizations should have overcome the initial challenges with AI, and so 2025 will be the year of letting it loose and seeing it bloom. — Jen Chew, VP Solutions & Consulting, Bristlecone

AI and Automation Will Take Over Tedious Vulnerability Management Tasks

Security teams are overwhelmed by the growing volume and complexity of vulnerabilities, leading to errors and burnout. AI-driven tools are set to change this, automating tasks like triage, validation, and patching. By analyzing vast datasets, these tools will predict which vulnerabilities are most likely to be exploited, allowing teams to focus on critical threats. By 2025, up to 60% of these tasks will be automated, significantly improving accuracy and response times. AI-driven tools will also proactively discover vulnerabilities, closing gaps before attackers can exploit them. — Jimmy Mesta, CTO and founder, RAD Security

AI Will Give CISOs and Security Teams a Head Start on Threats

It's no longer enough to detect threats after they've infiltrated a system. By training models on vast amounts of historical data, AI will help security teams spot emerging attack patterns before they cause damage. By detecting subtle anomalies in network traffic and user behavior, AI will provide proactive alerts, giving organizations a critical edge. This approach could cut the average time to detect threats (MTTD) by half. Moreover, as AI continues to advance, multi-agent systems will emerge as a new challenge. Attackers will use these systems to orchestrate sophisticated, automated attacks, forcing defenders to adopt similarly sophisticated AI solutions. — Jimmy Mesta, CTO and founder, RAD Security

AI Will Help Close the Cybersecurity Skills Gap

The demand for cybersecurity talent keeps growing, but there aren't enough skilled professionals to fill the gap. AI-powered tools are stepping in to level the playing field, helping organizations of all sizes automate threat detection, incident response, and compliance tasks. In the new year, over half of small and medium-sized businesses will depend on AI to manage their security operations. These tools will make advanced protection accessible, especially for teams with limited resources. — Jimmy Mesta, CTO and founder, RAD Security

AI-Driven Threat Detection Will Integrate Seamlessly into DevOps Workflows

AI will become fully integrated into DevOps workflows, enabling security to be embedded directly into the development process. With cloud-native environments growing more complex, AI-powered threat detection will continuously monitor applications in real-time, catching vulnerabilities before they can escalate. Rather than interrupting development cycles, AI tools will seamlessly provide proactive alerts and insights, helping teams address security issues as they arise — without slowing down the pace of innovation or deployment. — Jimmy Mesta, CTO and founder, RAD Security

AI Will Simplify Compliance in an Era of Stricter Regulations

As global data privacy and cybersecurity regulations become stricter, compliance will become an even more significant challenge. Traditional, manual compliance processes won't be enough anymore. By 2025, AI will automate compliance workflows, including auditing, reporting, and monitoring regulatory requirements in real-time. AI tools will identify gaps, generate actionable insights, and help organizations stay agile in adapting to evolving legal landscapes, freeing up security teams to focus on proactive protection. — Jimmy Mesta, CTO and founder, RAD Security

AI Workload Security Will Address New Attack Vectors

As AI becomes central to operations, attackers are targeting foundational elements like training datasets, where a single compromise can create widespread vulnerabilities. AI workload security will be crucial, focusing on protecting models from data poisoning, model evasion, and adversarial attacks. By 2025, integrated security solutions will safeguard AI throughout its lifecycle, ensuring data integrity and resistance to tampering. — Jimmy Mesta, CTO and founder, RAD Security

Agentic AI Will Transform into Agentic Workflows That Drive Exponential Efficiency and Innovation

As the adoption of agentic AI ramps up, we will also continue to extend it. In 2025, the technology will become mature enough to have multiple AI agents work together and feed into each other to orchestrate multi-step objectives — they will transform into agentic workflows that tie into each other to make decisions and perform more complex enterprise tasks. With agentic workflows, your systems will retain memory and intelligence and will have a high degree of adaptability in order to proactively adjust workflows based on the environment's responses. This multiplies efficiency and innovation exponentially. — Abhinav Puri, VP of Portfolio Solutions & Services, SUSE

For Executives, Optimizing AI Cost and Performance Will Require a Strategic Balancing Act

Firstly, identifying high-impact use cases will be crucial. This means prioritizing AI initiatives that directly contribute to core business objectives and offer measurable ROI, such as automating critical processes, enhancing customer experiences, or optimizing supply chains. Investing in robust data infrastructure and efficient AI models will also be key, ensuring the foundation for accurate and reliable AI-powered solutions. Secondly, embracing efficient AI practices will be essential. This includes leveraging solutions for scalability and cost-effectiveness, ensuring effective GPU utilization, fine-tuning AI models to reduce computational demands, and exploring techniques like model compression and knowledge distillation to optimize performance without sacrificing accuracy. By adopting a data-driven approach and continuously monitoring AI initiatives, executives can ensure they maximize the value of their AI investments while controlling costs. — Abhinav Puri, VP of Portfolio Solutions & Services, SUSE

Multi-modal AI Is Set to Revolutionize AI in 2025

Multi-modal AI will enable machines to process and integrate information from multiple sources like text, images, video and audio. This breakthrough will lead to more intuitive human-computer interaction, enabling us to communicate with AI seamlessly using voice, gestures, and visuals. Imagine AI assistants that understand complex requests involving multiple forms of media, or robots that can perceive and navigate their environment with human-like awareness. Furthermore, multi-modal AI will fuel a wave of innovation across industries. Expect personalized learning experiences that adapt to individual needs, AI-powered tools that revolutionize content creation by generating videos from text or music from images, and advancements in healthcare with AI analyzing diverse patient data for accurate diagnoses. However, this progress necessitates a focus on ethical considerations, ensuring fairness and responsible use of these capabilities. — Abhinav Puri, VP of Portfolio Solutions & Services, SUSE

Specialized Foundation Models Take Center Stage

Implementation complexity: full coffee IV drip needed; market readiness: initial adoption; investment required: significant investment. While large language models (LLM) have dominated the conversation, the real innovation is happening in specialized foundation models. Look at what's happening in drug discovery, material science, and agriculture. We've already seen models predict over 200 million protein structures and discover 2.2 million new materials. In 2025, this universe of models will accelerate. Every major player has their own language model — that's becoming table stakes. The true differentiators will be these domain-specific models tackling complex scientific and mathematical challenges. — Vijoy Pandey, SVP of Cisco's incubation and innovation engine Outshift

Show Me the Money: Agentic Apps Generate Revenue

CIO sleep loss: Regular midnight thoughts. Developer excitement: Clear whiteboard needed. VC funding frenzy: Term sheets flying. Right now, everyone's experimenting with AI agents, but nobody's making real money yet. That changes in 2025. The foundational technology is ready — but we still need to solve core challenges around data quality, operational costs, and building trust. We need better ways for agents to communicate and collaborate. Think about something as simple as agentic creation of market analysis for a product - sounds straightforward, but nobody has deployed it yet. The market is ready for practical solutions that can demonstrate clear business value, ROI, and trust. — Vijoy Pandey, SVP of Cisco's incubation and innovation engine Outshift

Agent Heterogeneity and the Sprawl Challenge

Complexity: Counting grains of sand. Stack Overflow questions: "Please help!" flood. Enterprise FOMO: "Quick, schedule a meeting." We're heading into a world of "agent heterogeneity" — different vendors, different capabilities, minimal standardization which will create a growing challenge: agent sprawl. As AI gets built into every application and service, organizations will find themselves managing hundreds or thousands of discrete agents. Without open standards and frameworks, this diversity creates chaos. It's like the early days of networking — we need common protocols and standards so these agents can discover, communicate, and collaborate with each other effectively. This standardization and interoperability will be essential for enterprises to effectively manage and scale their AI initiatives. — Vijoy Pandey, SVP of Cisco's incubation and innovation engine Outshift

From Solo Tasks to End-to-End Processes

Market readiness: Early experimentation. Industry disruption: Cross-industry transformation. Developer excitement: Keyboard literally smoking. Today's AI assistants are like solo performers — good at individual tasks like drafting emails or analyzing data. In 2025, we'll see the full orchestra — AI systems managing complex end-to-end business processes. Supply chains will be orchestrated by collaborating AI systems handling everything from demand forecasting to logistics optimization, all adapting in real time. The key is moving from isolated tasks to integrated workflows that deliver real business outcomes. — Vijoy Pandey, SVP of Cisco's incubation and innovation engine Outshift

AI and ML to Revolutionize Retail Supply Chains

As we move into 2025, AI and machine learning (ML) will reshape retail supply chains, driving efficiency and adaptability. More importantly, as the pace of product life cycles quickens, predictive analytics can help retailers anticipate shifts and restock faster, avoiding costly shortages or oversupply. From demand forecasting to personalized shopping experiences, technology is transforming retail at every touchpoint, enabling brands to build deeper connections and respond dynamically to consumer needs. — Keith Nealon, CEO, Bazaarvoice

AI-Powered Personalization

AI and machine learning are revolutionizing how brands engage with consumers. From personalized recommendations to automated customer service, these technologies offer insights and experiences at a scale that was previously impossible. And these experiences are what shoppers crave — according to our research, personalized offers drive 45% of shoppers to complete purchases online. In 2025, the brands that leverage AI to deliver hyper-personalized experiences and maintain a responsive, flexible supply chain will have a significant edge in building long-term customer loyalty. — Colin Bodell, Chief Technology Officer, Bazaarvoice

The Rise of Autonomous Agents

In 2024, we told AI what to do. In 2025, AI agents will start doing the actual work while we watch. Software engineers will input what they want, then see their screen come alive — the cursor moving on its own, opening files, writing code, running tests, fixing bugs. It's not about AI suggesting code in a chat box anymore. The cursor will literally move by itself, doing real development work. Microsoft and GitHub are already testing early versions. The change will be striking — from AI as a smart assistant to AI as a capable executor, handling full development workflows while engineers focus on higher-level decisions. — Andrew Feldman, Founder & CEO, Cerebras

The Emergence of 'Thoughtful' AI

Current AI is really just pattern matching — fast but shallow. 2025 brings something fundamentally different, sparked by OpenAI's O1 breakthrough in test-time computation. AI will start taking variable time to think through problems. Simple questions get instant answers. Complex system design questions? The AI will tell you "need a few minutes to think this through properly." It's a natural evolution — harder problems need more processing time. The implications are significant — AI that can tackle genuinely complex analytical work, taking more compute time when needed to generate better answers. — Andrew Feldman, Founder & CEO, Cerebras

The End of Nvidia's AI Chip Monopoly

The AI chip market is finally opening up beyond Nvidia's dominance. Cerebras is making real progress in high-performance inference, especially for companies running large language models. AWS and Google are rolling out chips optimized for cost-efficient AI workloads, while Apple is pushing AI computation to mobile devices. Nvidia will still sell every chip they make, but companies now have choices for different AI needs. The result? A more mature hardware ecosystem where different workloads can find their optimal chips — from edge devices to data centers. — Andrew Feldman, Founder & CEO, Cerebras

AI Crossed 3 Billion Active Users

AI is hitting growth curves we haven't seen since early mobile. We're at roughly 1 billion active users now — Meta's AI features reach 500M monthly actives, OpenAI has 300M weekly users, and that's not counting Google, Apple, and others. 2025 will see us cross 3 billion as AI becomes woven into the fabric of every major platform. The acceleration is driven by three forces: mobile AI getting serious with Apple and Android pushing on-device models, messaging apps making AI the default experience, and workplace tools embedding AI into daily workflows (think Microsoft Office, Google Workspace). When the core tools billions already use become AI-first, crossing 3B users isn't just possible — it's inevitable. — Andrew Feldman, Founder & CEO, Cerebras

AI/ML Won't Replace Engineers but Give Them Superpowers

While AI/ML continues to be an important emerging tool, we've moved beyond treating it as a catch-all solution. We've begun to hone in on specific use cases that deliver tangible value. AI/ML excels at pattern recognition, enabling the automation of time-consuming tasks and identifying cost-saving opportunities. Rather than replacing engineers, our solutions will augment their capabilities by reducing noise and providing well-reasoned recommendations. This frees up human experts to focus on complex decision-making where their expertise is most valuable. — Quynton Johnson, product marketing lead, Grafana Labs

Digital Memory Curation

In 2024, people were building reliance on generative AI tools like ChatGPT or Claude to answer questions and save time. In 2025, this will go a step further: AI will capture conversations and create insights to make you more productive. By finding patterns between conversations within meetings, calls, and videos, AI will start to establish a “digital memory” for its users. — Jason Chicola, CEO, Rev

AI Job Opportunities

AI will create a ton of new jobs just like the internet did years ago — some we can't even imagine yet. Positions like Prompt Engineer will start cropping up in 2025 as businesses focus more on AI ROI and push to see results. — Fernando Trueba, chief marketing officer, Rev

Industry-Specific LLMs and Consolidated Productivity Suites

In 2025, we'll see a shift toward industry-specific LLMs that are trained for specific sectors, and curated ecosystems of tools that integrate seamlessly across enterprises. These consolidated productivity suites will securely retain sensitive data and eliminate the need to repeatedly provide context, transforming AI from a novelty to essential business support. — Aron England, chief product and technology officer, Rev

A Bigger Focus on AI ROI

Companies spent an enormous amount of money on generative AI in 2024, but in many cases are still waiting to see its impact on top-line revenue. Although AI adoption has increased tremendously, we're still in the early stages where employees haven't quite mastered day-to-day use. In 2025, companies will draw a harsh line and say that if the AI tool isn't clearly contributing to ROI, it's gone. — Aron England, chief product and technology officer, Rev

The Rise of AI Multimodals

We'll see more quality maximalists — high-end models that can only run on a large server environment but can do it all. Single-purpose AI won't be as useful or prevalent as the dominant "omni" models start to emerge. For example, in the speech technology space, one model will take care of audio transcription, diarization, speaker labeling, and other needs in one fell swoop. — Lee Harris, VP of engineering, Rev

Powerful On-Device AI

No more sending your data to the cloud for outputs — more high-quality AI models will be available directly on mobile devices (phones, headsets, cars, etc.), and usable without Internet service or a huge server in the background. This will be a boon for legal, law enforcement, journalism, and other professions where real-time availability and data security are key. — Lee Harris, VP of engineering, Rev

AI Agent Cody Banks (just kidding)

2025 is set to be the year of the AI agent revolution. With exponential advancements in AI models and agentic workflows, AI agents will transform industries by automating complex tasks and enhancing efficiency. Major tech investments are accelerating this shift, making AI agents indispensable across business and daily life. The convergence of powerful technology and user-centric design will redefine productivity and innovation on a global scale. — Mike Diolosa, CTO, Qloo

Consumers Will Expect Greater Personalization as Apple Intelligence Raises Expectations

In 2024 we saw the rise of AI agents to assist humans in their day-to-day and enhance our efficiency. As AI agents become even more common, we'll start to see a rise in consumer demand for hyper-personalized experiences. At the same time, privacy-compliant consumer data will be essential for keeping on-device AI models truthful and valuable to end users. — Coby Santos, chief product officer, Qloo

Production Studios and Content Companies Will Outgrow Their AI Fears

Although 2024 saw controversy over the use of AI training data for content production, 2025 will see a boom in AI partnerships among big brands (such as Runway's recent announcement with Lionsgate). Companies are starting to see the benefits AI can provide alongside protective AI guardrails to safeguard their existing human talent and customer data, and as a result they'll start fighting for exclusivity with the biggest tech brands to enhance their content and efficiency. — Alex Elias, CEO, Qloo

Leaving Opt-out Defaults in 2024

Companies like LinkedIn and X have come under scrutiny for training AI models on consumer data by default, sometimes with no notification at all. Companies will revert to an opt-in strategy in 2025 and will ultimately protect their reputations by doing so, especially in a climate where consumer trust of AI is still not completely widespread. — Alex Elias, CEO, Qloo

More Personalized AI-Powered Travel Recommendations

AI-powered travel recommendations will grow more sophisticated and personalized, especially now that travelers prefer experiences that genuinely reflect their individual tastes and interests, instead of simply visiting the most popular destinations. On top of that, popular destinations are becoming undesirable as rising overtourism has created frustrated travelers and locals alike. — Jim Jansen, CRO, Qloo

Real Estate Developers to Turn to AI

AI is going to completely redefine what's considered prime real estate. Business districts in major cities are dying as empty office spaces and a slow worker return have culminated in a 52% drop in office value. In 2025, real estate developers will turn to AI to provide hyper-specific recommendations for restaurants and shops in these areas, turning these declining neighborhoods back into popular areas for consumers based on their unique taste profiles. — Levi Nitzberg, SVP of growth, Qloo

GenAI Hype Cycle Comes Back Down to Earth

Generative AI will never not be cool, but we reach a point where we give a slight nod to the hype cycle — and then get down to the business of delivering real value. This happens by simplifying our approaches, rules and models, complementing them with a targeted use of LLMs. Keep a close eye on that Nvidia stock. — Jared Peterson, Sr. VP, Platform Engineering, SAS

Impacts of AI Regulation

Regulation keeps AI in check but makes it challenging for businesses to use pure open source. Innovation takes a hit. Innovation silos crop up. AI enthusiasts cross their fingers that the impacts are temporary and hunker down to find solutions that work for their region. — Jared Peterson, Sr. VP, Platform Engineering, SAS

The Titans of Tomorrow Are AI Augmented Today

Fully AI-enabled organizations are the ones that will win the IT battles of 2025. As generative AI evolves from a "shiny new toy" to "just" another type of AI, organizations will fully operationalize AI to automate routine tasks that free employees for higher-value work. Those automations mean they'll make decisions faster, recognize opportunities more quickly, and drive more innovation than their competitors. In short: they'll win. — Jay Upchurch, chief information officer, SAS

LLMs Get Commoditized … and Specialized

In 2025, LLMs will become commoditized, leading to AI pricing models collapsing as base-level capabilities are offered for free. The real value will shift to specialized services and domain-specific applications built on top of these models. Simultaneously, the rise of open-source LLMs will challenge the dominance of a few key providers, driving a more decentralized AI landscape where customization and integration will be the key differentiators. — Udo Sglavo, VP, Applied AI & Modeling, R&D, SAS

The Future Won't Be Right for Organizations That Fail to Act on Generative AI

Think back to the digital transformation wave of the early 2000s. Companies that embraced the internet, digitized their processes, and invested in e-commerce became the Amazons, Googles, and Apples of today. Those organizations that waited or followed the wrong adoption path either adapted too late or disappeared entirely. Similarly, organizations that fail to act now will find it increasingly difficult to compete in the GenAI-powered economy. GenAI is not just another trend. It's the next leap in business evolution, and the organizations that understand this and move decisively will be the ones shaping the future. — Marinela Profi, global GenAI/AI strategy lead, SAS

The Semantic Layer Becomes the Enabler for LLMs in Enterprises

In 2025, the Semantic Layer will become the crucial enabler for LLMs in enterprises, acting as a bridge between internal data and LLMs to deliver precise, contextually relevant insights. By unifying enterprise data with global knowledge, this integration will revolutionize decision-making and productivity, making GenAI indispensable. Companies that embrace this convergence will dominate in innovation and customer experience, leaving competitors behind. — Ariel Katz, CEO, Sisense

Ethical and Secure AI Takes Center Stage

Companies will prioritize secure, customizable AI solutions that protect sensitive customer data while still leveraging the power of advanced analytics. AI governance frameworks will become essential for enterprises to ensure ethical use of AI in customer interactions and decision-making processes. Regulatory compliance in AI will drive innovation in transparent, explainable AI models for customer service applications. — Ashish Nagar, CEO of Level AI

LLMs Will Hallucinate Much Less

LLMs are still known to produce factually inaccurate or utterly random content, especially in languages other than English, where training data is sparse. There is a known phenomenon where the lexical coverage is sufficient to produce content in a foreign language. However, the actual cultural and social realia still stem from the English-speaking world or are entirely random. As the models are deterministic, they will produce an answer no matter what, even when the confidence level is low and despite prompts to respond with "I don't know." To combat this behavior, new techniques of detecting and mitigating hallucinations will continue evolving, opening the doors for more use cases where output reliability is vital. Such methods include analyzing correlations between edit rate, log probability, and semantic entropy, thus catching hallucinations and either following this analysis with a self-healing step, or sending potentially flawed content to a human-in-the-loop review; without a doubt, new approaches will be introduced both to prevent model hallucination and mitigate hallucinations in the post-processing step. — Olga Beregovaya, VP of AI, Smartling

Stand-Out Innovative AI Applications

The AI application that excites me the most is the idea of agentic AI — systems that can plan and execute tasks to meet goals that I define, assisting me in project planning, content creation, and ultimately making my life easier and or letting me focus on more creative tasks. — Jerod Johnson, Sr. technical evangelist, CData

Organizations Shift to Targeted AI Initiatives

Most organizations are moving towards AI-readiness. Recently I've noticed a narrowing of scope for AI projects — from an "AI will do everything" attitude to an "AI can help us/our customers in this very specific way." As orgs narrow their focus, they can create realistic goals for their AI initiatives and work backwards to determine how to build the models/training/etc. for their AI. A skill/infrastructure gap I see is effective access to data for every stakeholder in the organization. IT and developers have an easier time getting to data, because of their skillsets, but line of business users shouldn't be expected to know how to access data, while still being provided democratized access. — Jerod Johnson, Sr. technical evangelist, CData

More Content Will Shift to 'No Human in the Loop'

The quality of AI-generated content is getting exponentially higher, in many instances reaching human parity, especially for "not noisy," structured, professionally authored content types, like help systems, manuals, websites and eLearning content. About two or three years ago, the recommendation for such "branded" content would have been to use a "human in the loop" process. The global content transformation process is shifting towards instant, fully automated delivery. We will see more and more content types move towards "multilingual generation" (rather than traditional translation) using LLMs with RAG or few-shot examples and advanced prompt engineering. However, the adoption of this workflow will vary based on languages, simply due to the availability of model training data. Under-resourced languages will still require significant human effort for the generated or translated content to be at the level of quality needed. — Olga Beregovaya, VP of AI, Smartling

Governance, Legal Frameworks, and Ethical Considerations Around AI Will Be More Structured and Transparent

We could label 2023 as a year of "Generative AI Chaos," where there were more questions than answers when implementing AI-based technologies. The Infosec questionnaires and corporate or government guidelines were rather vague, and there was a lot of uncertainty about IP, data protection, PII handling and overall risk assessment. 2024 became a year of "measured deployment," where the learnings of AI implementations were being translated into standards and regulations. There are two aspects to such regulations: the ethics of actual deployments, where potential impact is analyzed and risks are mitigated, and the ethics of ensuring safety and emotional well-being of the workforce. We already see more local governmental regulatory bodies' initiatives around safe AI and such initiatives will have more global alignment in the future. — Olga Beregovaya, VP of AI, Smartling

2025 Sees AI Governance Surge

New standards drive ethical, transparent, and accountable AI practices: In 2025, businesses will face escalating demands for AI governance and compliance, with frameworks like the EU AI Act setting the pace for global standards. Compliance with emerging benchmarks such as ISO 42001 will become crucial as organizations are tasked with managing AI risks, eliminating bias, and upholding public trust. This shift will require companies to adopt rigorous frameworks for AI risk management, ensuring transparency and accountability in AI-driven decision-making. Regulatory pressures, particularly in high-stakes sectors, will introduce penalties for non-compliance, compelling firms to showcase robust, ethical, and secure AI practices. — Luke Dash, CEO, ISMS.online

Generative AI in Mainstream Creative Production

By 2025, generative AI is poised to revolutionize mainstream media, with film studios, music producers, and massive brands harnessing its power to create hyper-realistic visual effects, original soundtracks, and high-quality content at unprecedented speeds. The era of human-AI collaboration will blur the line between artist and algorithm, sparking a new wave of creative output that enhances productivity and imagination. — Andy Lin, CEO, Provoke Solutions

Personalized Learning and Education Tools

Generative AI will be a game-changer in education, enabling platforms to offer hyper-personalized learning experiences tailored to each student's unique needs. With adaptive content generation and real-time feedback, students will benefit from curriculums that adjust dynamically, fostering deeper engagement and accelerating learning. — Andy Lin, CEO, Provoke Solutions

Enhanced Synthetic Media Regulation and Ethics

As generative AI becomes more pervasive, robust ethical guidelines and content verification methods are becoming increasingly important. By 2025, we can expect the establishment of global standards and regulatory bodies focused on preventing misuse and ensuring responsible use of AI — tools for identifying deepfakes and ensuring the adoption of content authenticity across digital platforms. — Andy Lin, CEO, Provoke Solutions

Advancements in AI-Generated Human Interaction

Chatbots and virtual assistants powered by generative AI will evolve into entities capable of sophisticated, empathetic communication. By 2025, these AI models will be able to conduct nuanced conversations that mirror human interaction more closely than ever, transforming customer service, therapy, and personal digital companions into highly effective, human-like experiences. — Andy Lin, CEO, Provoke Solutions

Democratization of Generative AI

Generative AI will become more accessible, with tools designed for individuals and small businesses becoming as ubiquitous as office software. This democratization will empower users to create professional-grade content and prototypes without extensive technical knowledge, fostering innovation and entrepreneurship on a global scale. — Andy Lin, CEO, Provoke Solutions

AI Agents Will Become Our New Coworkers, Ushering in a New Era of Hybrid Human-Tech Workforces

Despite all of the advancements AI agents have seen in 2024 — taking on more complex, multi-step tasks organization-wide — we've barely scratched the surface of what these agents will accomplish in the coming year as AI chips become more powerful. These autonomous agents will become the middlemen between humans and their tech stacks — equipped with the ability to speak the language of both — streamlining processes and creating entirely new ones that will drive a never-before-seen level of productivity and innovation. Organizations will also find new ways to integrate their AI colleagues into everyday workflows to take advantage of the exponential efficiency gains agentic AI will create. — Marco Santos, co-CEO, GFT

Companies Will Prioritize Readying Their Business for AI Before Implementing More AI Tools

Many companies implemented AI into their business in 2024 — but not every company was ready for AI. Companies have quickly learned that implementing AI alone is not enough; they need to build an AI-ready culture across their organization too. In 2025, we'll see companies that were quick to hop on the AI train slow down and prioritize readying their organization for AI before adding new technology to their stack. This means offering employees the proper training and education to use AI to its full potential. Currently, only 23% of employees feel completely educated and trained on AI. Companies will also need to ready their data for AI by centralizing what's spread across multiple systems, channels and gatekeepers. By bringing together all of a company's data in one place, AI can provide insights and recommendations that are based on a holistic view of the organization and projects — and will drive better business decisions. The companies that focus on readying their organization for AI before introducing more tools and technology will be those that see the most benefit from AI in 2025. — Dean Guida, CEO of Infragistics and founder of Slingshot

AI Regulations Will Continue to Become More Complex

While AI presents exciting opportunities, organizations must evolve their governance, risk, and compliance (GRC) strategies to address emerging privacy and security challenges. The good news is that many aspects of AI risk assessment can build upon existing GRC practices — organizations should start by applying traditional due diligence processes for systems handling sensitive data, then layer on AI-specific considerations like model drift monitoring and explainability requirements. Success in 2025 will depend on developing comprehensive frameworks that combine robust security fundamentals with AI-specific controls, particularly as regulations continue to evolve around high-risk AI systems. — Kyle McLaughlin, general counsel, Secureframe

AI Will Scale Efforts That Previously Required a Highly Specialized Workforce

As AI capabilities expand, Managed Service Providers (MSPs) are becoming even more critical in helping organizations navigate this complex landscape. While AI can now automate many traditionally labor-intensive tasks like patch management and vulnerability scanning, MSPs bring the strategic expertise needed to properly implement and govern these AI systems. The most successful MSPs in 2025 will be those who position themselves not just as technology implementers, but as strategic advisors who can help clients balance AI automation with proper security controls and human oversight. — Aaron Melear, VP Partnerships, Secureframe

GenAI Will Be Evaluated on ROI, and Achieving It Will Be Through Cost Control

To maximize ROI, you can either reduce costs or increase returns. I think the former is lower-hanging fruit in the AI space. 2025 will be a year when top executives begin to lose appetite for experimentation and innovation stories and start to look down more quantitatively at the AI commitments and investments that they've made over the last couple of years. They will start demanding their teams and agencies to prove returns. A much easier valve to release some of that pressure will be on the cost-optimization side. I think there will be fortunes made by those who figure out how to wield AI effectively while reducing its overall cost, footprint, and overhead. The irony is that the AI itself is probably the best weapon in reducing its own cost, as people start to delegate the data modeling, training, and querying to AI agents, thus reducing the volumetric consumption and compute on these AI models that currently cost enterprises so much. — Greg Brunk, head of product and co-founder, MetaRouter

Agentic AI Will Make Waves in 2025

You'll be hearing a lot about agentic AI in 2025. That's because agentic AI marks a genuine and significant departure from previous iterations of AI. Unlike earlier generative AI models, which primarily executed linear prompts (e.g., "do X"), agentic AI can "think" through sequences and execute multi-step problems in a highly logical order (e.g., "do X while considering the ramifications of Y and balancing for Z"). Another distinctive feature of agentic AI is its ability to "converse" with other tools and solutions, triggering actions across various platforms. This makes it far more dynamic and capable of orchestrating complex processes. As advancements continue and the most forward-thinking business leaders start deploying these solutions, expect much of the AI conversation in the coming months to shift to agentic AI. — Vall Herard, CEO, Saifr

AI Is Going to Become More Thoughtful in 2025

When we prompt a large language model (LLM) like ChatGPT or Claude to complete a task, we typically expect a response within 60 seconds. This quick, conversational style has become emblematic of LLMs in general — but that may soon change. Recent advances in AI, particularly breakthroughs in agentic AI, will lead to longer "thinking" times for complex queries as these models strive to deliver more context-rich, nuanced outputs. Agentic AI differs from earlier manifestations of AI due to its ability to process complex, multi-step problems in a non-linear fashion. This allows it to handle more sophisticated tasks and interact with multiple tools simultaneously, much like a human might. While impressive, this capability requires more processing power. AI practitioners should expect a "splitting" of AI as more complex tasks start requiring longer processing times, while simpler prompts continue to receive lightning-fast responses. Thus, leading LLMs will become far more advanced in 2025, but they may also take longer to provide good outputs, prioritizing depth and precision over speed. — Arindam Paul, vice president, Data Science, Saifr

Single Most Important Tech Trend in 2025: GenAI, Of Course

While it's probably not much of a surprise, I anticipate generative AI will continue to dominate the tech landscape in 2025. In 2025, tech vendors will start to move away from placing importance on standalone AI products and look more to the bigger picture of what this technology can do, especially regarding unstructured data. As generative AI progresses, so do its applications for synthesizing insights, enabling better access to information, and enhancing decision-making. This shift will move us beyond the confines of standard AI products, and instead, tech vendors will seek out ways to serve customers more holistically and outside the box of just individual solutions. — Steve Watt, CIO, Hyland

Enterprises Will Opt for Small Purpose-Built LLMs

Small, purpose-built LLMs will address specific generative AI and agentic AI use cases, powered by retrieval-augmented generation (RAG) and vector database capabilities. The number of both generative and agentic AI use cases will expand, and the need for ultra-low latency inference will increase, pushing more and varied AI models to edge environments. — Kevin Cochrane, CMO, Vultr

Silicon Diversity Will Revolutionize AI Efficacy/ROI

A broader array of state-of-the-art GPUs will enable purpose-built AI models to drive the next wave of innovation in the enterprise. 2025 will see increased attention to matching AI workloads with optimal compute resources, driving exponential demand for specialized GPUs. Silicon diversity — the emergence of highly specialized AI compute chips — will provide tailored solutions for each stage of the AI model lifecycle. Organizations that embrace this diversity will enjoy enhanced AI capabilities at reduced costs. Those who fail to leverage silicon diversity will risk falling behind in both performance and cost efficiency. — Kevin Cochrane, CMO, Vultr

We Will Witness the Great Rebuilding of Enterprise in 2025

Since ChatGPT burst onto the scene in 2022, GenAI has been the undisputed star of the AI Era. Now, GenAI is about to become the backbone of enterprise technology. As businesses have determined where AI fits into their operations and how to maximize its value, they are moving out of an adoption phase and into a reconstruction phase. Enterprises are now rebuilding their business operations with generative AI at the core, which will kick off an era of radical transformation in productivity and operational efficiency in 2025. — J.J. Kardwell, CEO, Vultr

AI Pricing Will Shift to More Competitive, Value-Based Pricing Models

Similar to how we saw SaaS drive subscription-based pricing models, AI pricing will shift to more competitive, value-based pricing models in 2025, such as usage-based or outcome-based pricing. When it comes to AI use cases and applications, total-cost-of-ownership (TCO) remains at the top of business leaders' minds. This means IT leaders are responsible for demonstrating AI's value to the business. In turn, the pressure is put on AI providers to ensure their pricing model matches the value they're offering, particularly regarding business outcomes. Value-based pricing models, such as usage-based pricing and outcome-based pricing, will become more prominent for AI providers to remain profitable without deterring customers simply because they don't see enough value to justify the cost. — Frederic Miskawi, VP and AI Innovation Expert Services lead, CGI

GenAI Grows Up — Real Use Cases to Keep It in Check

By 2025, the initial hype around generative AI, ignited by ChatGPT, will level off as the technology matures. These tools will evolve to meet the specific needs of business users more effectively. Foundation models will continue advancing, delivering increasingly accurate and relevant responses. Meanwhile, the broader market will develop the necessary infrastructure and tools to integrate GenAI into everyday operations. We'll see the emergence of new guardrails, enhanced techniques for building trust, and a broader range of use cases beyond chatbots. More "AI Agents" will be deployed to automate business processes and deliver insights directly to users. Consequently, platforms offering governance over GenAI and the tools to create agents will become increasingly important. — Christian Buckner, SVP, analytics and IoT, Altair

Geometric Deep Learning Will Transform Engineering Design Landscape

While generative AI has made waves across industries, a new technological advancement is poised to revolutionize engineering: geometric deep learning. This cutting-edge AI approach builds on recent breakthroughs in integrating AI with simulation software to accelerate decision-making and product development cycles. Geometric deep learning takes these advancements a step further by training machine learning models using existing simulation data and learn about 3D shapes at a level of understanding comparable to human perception of everyday objects. This capability dramatically speeds up design decisions to 1,000 times faster than conventional methods while expanding the boundaries of innovation. — Fatma Kocer, VP, engineering data Science, Altair

AI Chips War: Speed to Innovation Is the Winning Formula

As AI chip demand continues to surge, semiconductor companies will realize the critical role emerging technologies play in the design process. By integrating AI with simulation software, engineers can test new concepts and make design decisions up to 1,000 times faster than traditional methods, dramatically speeding time to market and cutting costs. This approach will be key to producing high-performance chips more efficiently and staying competitive in the rapidly evolving semiconductor industry. — Sarmad Khemmoro, SVP of technical strategy, electronics design and simulation, Altair

Scaling Models, Shrinking Resources

The world is realizing that simply scaling AI models without regard to efficiency isn't sustainable. In short, the era of one-size-fits-all models is ending. In 2025, expect the rise of "right-sized," industry-specific, AI: models designed to maximize impact with the lowest possible resource footprint. Companies will shift from "bigger is better" to "smarter is better," with a focus on hyper-customized small language models, tailored to their specific industries, and breakthroughs in AI efficiency that drive competition. Think fewer GPUs, more results. — William Falcon, founder and CEO, Lightning AI

LMs Evolve from Chatbots to Business Partners

Language models will move beyond chat interfaces to become more integrated into the business decision-making process. Imagine an LM in finance conducting due diligence in seconds. Language models will play pivotal roles, performing nuanced analyses that once required expert human intervention. — Luca Antiga, CTO, Lightning AI

Ethical AI Becomes Mission-Critical

2024 will be the year ethical AI moves from "important" to "essential." We're seeing a new wave of regulatory frameworks around AI transparency and fairness, and companies are feeling the pressure to adopt bias detection and transparency by design. This trend will extend into partnerships, acquisitions, and hiring, where an ethical AI strategy will be a prerequisite for market success. — Priya Shivakumar, COO, Lightning AI

Generative AI in Creative Industries

Generative AI isn't just supporting creatives — it's reshaping their fields. From film to advertising, generative tools are being used to prototype, brainstorm, and create, turning human-AI collaboration into an art form in its own right. Yet as these tools expand, human oversight will ensure that AI creations reflect the unique cultural, ethical, and artistic values that only people bring. — Priya Shivakumar, COO, Lightning AI

Organizations Will Be More Cautious of 'AI Washing'

If 2023 was the year of AI experimentation, 2024 would be the year of hands-on AI implementation, and 2025 will be the year of clarity. Organizations will have an increased focus on actual ROI of the outcomes of AI systems adoption, in response to the rising concern of AI washing. In 2024, the SEC confirmed they'd continue to closely examine all AI-related claims made by companies or firms, whether public or private, seeking to attract investors and raise capital. It will be crucial for organizations to measure AI's success and costs accurately, providing as much clarity as possible. Clarity must be constructed around regulations and frameworks, as well as the value each organization hopes to garner from AI, to ensure maximum AI ROI and reduce any association with AI washing. — Phil Lim, AI champion, Diligent

AI Governance Will Become Even More Critical

AI Governance Frameworks will mature but remain largely unadopted with no clear industry standardization in 2025. Risk-averse organizations will shun AI completely, and organizations with high-risk appetite will continue to "move fast and break things," setting both types of organizations up to fall behind. Organizations best equipped to navigate the uncertainty with strong AI governance will be the most successful. This includes implementing the proper process, technology, and people, like a Chief AI Officer or similar role, to ensure responsible use of AI and help bridge the knowledge gap between leadership and the faster-adopting line-staff. — Phil Lim, AI champion, Diligent

Expect a Period of Mass Consolidation of 'AI-Native' Companies

FOMO and falling interest rates will spur consolidation as traditional companies gobble up AI start-ups who start to run out of cash. This will be a massive culture clash and very few organizations will come out ahead, but those that do will win big. A lack of education and understanding of AI, its limitations, and poor governance will lead many organizations to over-invest in dubious promises lacing a long-term competitive advantage. By 2030, we can expect the majority of the consolidation to be complete. — Phil Lim, AI champion, Diligent

Finding Real ROI in AI — Efficiency, Edge, and Cost Management

2025 will usher in a more measured approach to AI investment, as organizations will be increasingly focused on quantifiable ROI. While AI can deliver immense value, its high operational costs and resource demands mean that companies need to be more selective with their AI projects. Many enterprises will find that running data-heavy applications, especially at scale, requires not just investment but careful cost management. Edge data management will be a critical component, helping businesses to optimize data flow and control expenses associated with AI. For organizations keen on balancing innovation with budgetary constraints, cost efficiency will drive AI adoption. Enterprises will focus on using AI strategically, ensuring that every AI initiative is justified by clear, measurable returns. In 2025, we'll see businesses embrace AI not only for its transformative potential but for how effectively it can deliver sustained, tangible value in an environment where budgets continue to be tightly scrutinized. — Nick Burling, senior vice president, Product, Nasuni 

'Agent' Is the Term You Need to Know for Decision Making in 2025

AI agents will redefine how businesses and industries tackle complexity, autonomously leveraging vast datasets and executing strategies faster than human teams ever could. From marketing budgets to supply chains, AI agents will become embedded in critical decision-making processes, driving competitive advantage for those who embrace them early and properly. While their potential to improve operations and unlock new efficiencies is incredible, any reliance on a black box for decision making created significant challenges in accountability and trust, as well as determining solution quality. — Jerry Yurchisin, data science strategist, Gurobi Optimization

True Competitive Edge Will Belong to Organizations That Fully Integrate, Orchestrate AI into Daily Workflows

In 2025, forward-thinking organizations will shift from treating AI as an isolated solution for specific tasks. Instead, organizations will integrate AI across the entire enterprise, driving value-added outcomes and cohesion throughout every department and process. This will be a step change from executives "just throwing in AI" out of fear of missing out, to being able to maintain, adapt, and evolve otherwise isolated point solutions. While point solutions can bring short-term gains, they often create technical debt and complexity that stalls innovation in the long run — a value trap that ultimately makes future AI adoption even harder. Over the next year, as AI continues to generate excitement, businesses must look past the hype and thoughtfully consider how it can really benefit customers lastingly. Organizations need to orchestrate AI like any other endpoint within their end-to-end automated business processes to get the maximum benefits from their AI investments. The shift from adopting AI incrementally to fully integrating it will result in better and faster business outcomes, more adaptive business strategies, and a new level of agility that will set industry leaders apart from the rest. — Daniel Meyer, CTO, Camunda

The Rise of Agentic AI in API Security

With the growing use of agentic AI, where bots act autonomously on behalf of users, traditional methods of distinguishing malicious automated activity will become obsolete. Security systems will shift focus from detecting automation to predicting behavior and intent, introducing a new frontier of challenges in API security & Bot Management. — Will Glazier, director of Threat Research, Cequence Security

AI Moves from Hype to Business Outcomes

Brands need to stop spending millions on AI experiments that never make it out of the lab. In 2025, brands will demand to see business outcomes before committing any investment to AI. They'll seek out and invest in solutions that deliver real results in weeks, not months or years. Baseless claims and AI hype will be largely ignored as brands focus on vendors that can deliver measurable AI business outcomes. — Dave Singer, global vice president, Verint

Ungoverned GenAI Will Wreak Havoc

When we first saw the capabilities of generative AI, businesses were astounded at the possibilities. The CX industry looked forward to a future where we'd never have to write any content and every customer question would magically be answered in the blink of an eye. Now reality has set in. GenAI hallucinations, bias, privacy concerns, and more have caused countless problems for brands deploying AI without any guardrails. 2025 will see contact centers take a smarter approach to generative AI, with defined governance to ensure accurate results while mitigating risk. — Dave Singer, global vice president, Verint

AI's Business Impact Takes Shape in 2025

As tech vendors added AI capabilities to their offerings in 2024, we have seen IT departments adopt and release these features. This has resulted in small to moderate business improvements, but not the seismic shift that might have been implied by the hype. Leaders are still trying to crack the code of how AI will bring either significant productivity, reduction in operational complexity, or improvements in employee experience. 2025 will start to show more meaningful progress towards using AI to create business impact. First, AI will show value for employees to "shift left" and perform tasks of greater value and complexity, while digital AI agents will be able to answer the typical questions handled by front line support teams. Bots will be ubiquitous to all organizations to answer inquiries ranging from Sales account summaries to HR benefits. Second, the introduction of reasoning capabilities will ultimately be a game changer, but each organization will need to assess how to utilize it. 2025 will mark the beginning, but not the end, of that journey. — Mindy Lieberman, CIO, MongoDB

A More Thoughtful Approach to AI Adoption

In 2025, we can expect the focus to shift from "what AI can do" to "what AI should do," moving beyond the hype to a clearer understanding of where AI can provide real value and where human judgment is still irreplaceable. As we advance, I think we'll see organizations begin to adopt more selective, careful applications of AI, particularly in areas where stakes are high, such as healthcare, finance, and public safety. A refined approach to AI development will be essential — not only for producing quality results but also to build trust, ensuring these tools genuinely support human goals rather than undermining them. — Tara Hernandez, VP of developer productivity, MongoDB

Multi-Modal Training Will Become More Mainstream

In 2025, multi-modal training, which integrates different types of data — such as text, images, audio, and video — will become a more dominant approach in model training. This shift is driven by the need for AI systems to better understand and process the complexity of real-world data, allowing for richer and more context-aware applications. For example, multi-modal models can improve use cases like autonomous driving, where understanding visual, auditory, and textual information is critical. The rise of these models will also spur demand for more advanced hardware and storage solutions, as the complexity of training environments continues to grow. — Haoyuan Li, founder and CEO, Alluxio

Pre-Training Will Become a Key Differentiator for Organizations Adopting LLMs

By 2025, pre-training will emerge as a crucial differentiator among organizations developing large language models (LLMs). As the AI landscape evolves, access to vast amounts of high-quality data — especially industry-specific data — will become a major competitive advantage. Companies that can effectively harness big data infrastructure to leverage their large-scale datasets will be better positioned to fine-tune their models and deliver more effective, specialized solutions. However, this also introduces a significant bottleneck. Preparing and curating the right data for pre-training is increasingly complex, and companies without robust big data infrastructure will struggle to keep up. Efficiently handling this data preparation, cleaning, and transformation process will become a critical challenge in the race to develop more powerful and relevant LLMs. — Haoyuan Li, founder and CEO, Alluxio

Agentic AI Takes on Front-Line Interactions

In 2025, agentic AI will transform customer service and sales by autonomously handling routine front-line interactions, freeing human teams to focus on higher-level tasks requiring empathy, creativity, and strategy. These AI agents, working alongside human users but capable of autonomous decision-making, will manage a range of customer touchpoints — from responding to inquiries and scheduling appointments to troubleshooting and qualifying leads. Building on predictive and generative AI, agentic AI complements these capabilities to create more intelligent and productive interactions. In customer service, it will handle common inquiries, troubleshoot efficiently, and continuously learn from interactions to enhance responses, delivering a more tailored, dynamic user experience. In sales, agentic AI will autonomously qualify leads, track customer behaviors, and prioritize follow-ups, ensuring that sales teams engage only with the most promising prospects. This expanded capability not only boosts productivity but also shifts human roles toward strategic, relationship-driven work, where creativity and empathy shine. The result is a powerful synergy between humans and AI, fostering smarter customer engagement, higher satisfaction, and a more empowered workforce. — Burley Kawasaki, global VP of product marketing and strategy, Creatio

AI Assistants Become In-Car Companions

In 2025, more manufacturers will adopt AI into their vehicles, specifically bringing generative AI into the car with assistants, graphic and music generation. But in order to become a true companion inside the car, AI needs to understand in-cabin occupants completely and become familiar with all modalities that humans use. Emotion AI and advanced sensing technologies can enhance the in-cabin experience with generative AI by revealing the emotions, distraction levels, body and facial behaviors, and the visual and audible interaction between drivers and passengers. — Detlef Wilke, VP of Innovation & Strategic Partnerships, Smart Eye

The Key to AI in Advertising? Humans

Almost everywhere you look, industries are AI obsessed — but marketers and advertisers have been leveraging AI for years. However, AI is not the answer to captivating and creative ad campaigns. In fact, AI-only-generated ads often produce strange outputs, resulting in choppy, loud, disjointed spots. The ad industry will always rely on people during the creative process and testing stages to create ads that evoke emotions and connect with consumers in meaningful ways. — Graham Page, global managing director, media analytics, Affectiva

AI Is Entering a New Season

The introduction of ChatGPT two years ago sparked an "AI summer" of massive excitement and investment. According to a recent CNBC report, $26.8 billion was invested in nearly 500 generative AI deals, extending the 2023 trend when GenAI companies raised $25.9 billion. The "bubble" is unlikely to burst in 2025, but we are entering an "AI fall" as organizations struggle to scale the implementation of AI and where investors, business leaders and boards start expecting returns on their investments. This adjustment will likely lead to a year-over-year pullback in funding for GenAI startups and a further concentration of funding on the select few startups that are getting market traction. The expectation will not be net job losses from AI in advanced economies, and in the U.S. it could actually drive job creation as companies seek to meet demand for tailored AI solutions that fulfill specific business use cases. However, developing countries with large customer service and back office-processing industries are likely to see significant job loss in part due to AI. — Kjell Carlsson, head of AI strategy, Domino Data Lab

AI Will Become Boring (That's a Good Thing)

Generative AI has been the shiny exciting object for two years now, but that's about to change. There has been a lot of buzz around its capabilities and the massive investment flowing into GenAI startups. However, AI is already starting to transform from being that shiny new toy that automatically solves everything workers struggle to do — to being just another technology that solves targeted problems, requires hard work, extensive skills, and specialized capabilities to deploy. In short, AI will transition from being amazing and impractical, to being boring and impactful. Agentic AI will continue to be hyped as the next big thing that will replace the majority of human tasks, but that won't happen in 2025, if ever. Instead, organizations are starting to set their sights on a more practical variant of agentic AI where AI automates narrow, highly controlled tasks — like getting a rebate on your delayed food delivery. In short, the hottest part of AI in 2025 will be the boring, but valuable, topic of AI Engineering — how to integrate, operationalize and govern the ecosystem of technology components you need to make AI solutions work. — Kjell Carlsson, head of AI strategy, Domino Data Lab

The Age of 'Decision-Making Machines'

Generative AI will move beyond content generation to become the decision-making engine behind countless business processes, in everything from HR to marketing. IDC predicts that by 2025, 30% of major brands will be generating at least 50% of their ad copy using GenAI, but the real power will be in AI-driven business decisions, not just content. — Ravi Ithal, GVP and CTO, Proofpoint DSPM Group

Techniques to Properly Capture Evolving AI Capabilities

Fine-tuning LLMs and other AI models with proprietary data using techniques like retrieval-augmented generation (RAG) is a critical recent advancement that allows the fast-growing capabilities of these AI models to be brought to bear on tasks specific to the business. The biggest challenge in 2025 will be identifying and aggregating all this disparate unstructured data. Tools for managing structured data are much more mature in comparison. — Michael Allen, CTO, Laserfiche

Cyber Budgets Will Remain Flat & GenAI Assistants Will Form Cliques

Cybersecurity budgets will hold steady in 2025, but GenAI will prove more effective than ever. Cybersecurity programs are poised to grow with GenAI by boosting operational efficiency, reducing time-intensive tasks, and empowering businesses to do more with less. This will be particularly evident for industries with narrow margins, such as smaller manufacturing and healthcare companies that will continue operating under tight regulations. For these organizations, improving operational efficiency could make all the difference in reducing risk, given its ability to sift through massive volumes of data, identifying anomalies and risks faster than traditional methods. Additionally, GenAI models will become more specialized—tuned to the unique needs of specific industries—allowing ease of adoption and deployment. Ultimately, GenAI's role will go beyond driving efficiency, to transforming how security teams operate by shifting resources from reactive to proactive approaches. In 2025, we'll see GenAI not only reduce workloads but also drive strategic decision-making, making cybersecurity a true enabler of growth and resilience. — Gaurav Banga, CEO and founder, Balbix

Mad Scramble for AI Guidelines and Frameworks

With GenAI tools now ubiquitous, 2025 will see a frantic scramble to rein in AI — just as we saw with social media. The focus will not only be on protecting users but also on having frameworks to safeguard AI from other AI. Frameworks and guidelines will be pushed at three levels: international (e.g. the EU), regional (e.g. NCSC), and organizational. The organizational level will likely be most effective due to clear guidelines on acceptable use and security, while higher levels become less effective. International regulations often allow room for interpretation, enabling businesses to circumvent them. — Michael Adjei, director, Systems Engineering, Illumio

Custom AI Will Lose Favor Over Tried-and-True Tools

In the race to AI, many are pushing companies to create their own bespoke stack or infrastructure. There's no doubt the largest merchants (Amazon, Walmart, etc.) benefit from having their own black-box engine rooms, but for most, this is all planning and little profit. In 2025, we're going to see less experimentation with custom AI and more adoption of tried-and-true AI tools to optimize the shopper experience. Merchants will use internal and customer-facing AI tools to increase productivity, and they'll be transparent with shoppers about how it's helping them in their journey. — Zohar Gilad, AI expert and CEO/founder, Fast Simon

Proactive AI

Since ChatGPT arrived on the scene, CIOs have been experimenting rapidly with AI across a wide range of areas. In 2025, there will be a reckoning that forces them to focus on areas where those investments truly pay off. AI is not right for every use case and budgets are limited. That means CIOs will need to look hard at what their employees are trying to achieve and determine where the costs are truly justified by improvements in productivity and efficiency. That means we will likely see investments in areas like agentic workflows that improve customer service, and in AI tools and services that reduce friction in the workplace and make the employee experience more rewarding. — Faisal Masud, president, HP Digital Services

The End of the AI Honeymoon

Although the vast majority of organizations still plan to invest in AI next year, leaders now face pressure to show tangible business value.  With traditional data systems overwhelmed by growth in sensor, IoT, and network data, Predictive AI and machine learning will become essential for cost-effective, data-driven decision-making. Enterprises will now measure AI by its direct impact on business goals. Those that leverage predictive technologies to streamline insights and drive efficiency will lead in the new era where AI's worth is defined not only by innovation alone but by meaningful results on the bottom line. — Chris Gladwin, CEO and founder, Ocient

The Rise of AI-Powered Customer Support

In 2025, traditional chatbot experiences will fall short as customer expectations evolve. Businesses will increasingly adopt autonomous, AI-powered agents capable of delivering more adaptable, responsive, and personalized support. Generative AI, especially through conversational AI copilots, will enhance both customer and agent interactions by enabling faster, more insightful responses that feel human. While this shift opens vast opportunities, it also brings challenges in responsible AI implementation. As organizations scale up AI adoption, they'll need to establish guardrails, ensure transparency, and focus on regulatory compliance. Ethical considerations and transparency in AI decision-making will be essential to building customer trust. — Aurélien Caye, solution specialist, Sprinklr

Optimizing AI Efficiency and ROI in 2025

As organizations move beyond the initial generative AI hype, 2025 will see a focus on optimizing AI model efficiency. Companies will prioritize "smaller LLMs" or open-source, in-house models to improve ROI and manage costs effectively. Multi-modal capabilities in AI will gain traction, allowing systems to interpret diverse content formats and provide more comprehensive support. Success will come from blending AI's capabilities with human input to create meaningful customer experiences, ensuring that AI-driven transformations remain both sustainable and valuable. — Aurélien Caye, solution specialist, Sprinklr

AI Regulations

As with any emerging technology, regulation often lags behind rapid developments. We're already seeing many organizations implementing dedicated AI policies to evaluate and control the AI services they use. Current initiatives primarily focus on data privacy and the potential for AI to make critical errors. Initially, we can expect AI safety standards to evolve and become integrated into existing frameworks or form independent standards. Regulation may then extend to ethical considerations, defining acceptable versus unlawful uses of AI. Another significant question concerns legal responsibility: if an AI tool enables someone to commit a crime, does the tool's provider share liability? AI presents complex challenges for regulators, who must balance its potential benefits with its inherent risks. — David Kellerman, field CTO, Cymulate

Mastering AI Communication

AI is an emerging tool that has the opportunity to change how we interact with technology. Creating a basic RAG system with a LLM behind it to access specific information and convey it like a human is getting easier by the month. As AI grows in popularity, I believe people will need to learn how to properly talk to an AI similar to how people have to learn the best way to Google something. It can also change the learning process by helping with a more dynamic way of learning, capable of clarifying information and correcting mistakes without the need of human intervention. Easily the most important application for AI will be the introduction of RAG (retrieval augmented generation) to assist in customer support, documentation searching, and search engines in general. Being able to properly use these tools within an organization will be an extremely valuable upskilling organizations will need to take. — Josh Meier, generative AI author, Pluralsight

AI Has Been Such a Buzzword in Insurtech That It Is Difficult to Grasp What Is Real and What Is Hype

For the last five years, several of our carrier and product partners have mentioned their efforts to leverage AI in the underwriting process, but very few insurance or benefit companies have considered the amazing impacts AI could have on sales, adoption, and broader utilization of insurance and benefit products themselves. This is why Genius Avenue focuses on using conversational AI throughout the user journey and developing our platform to integrate with best-of-breed solutions. Having an AI "agent" online 24/7 to translate insurance policies into intelligible, human language and present the actual benefits and limitations, could have a transformational impact on the user experience. — Megan Wood, president, Genius Avenue

Filling Visibility Gaps Will Drive GenAI Data Platform Growth

Although the technology for GenAI's data ecosystem exists, deployment remains inconsistent. In 2025, enterprises will focus on filling visibility gaps by enhancing their platforms to support vector data, similarity search, knowledge graphs, and raw data stores. This will require balancing data control with accessibility while integrating GenAI into core systems for better insights and control. As enterprises scale from trials to full deployment, their systems will face new challenges. To unlock GenAI's full potential, platforms must handle massive data ingestion and provide parallelized access to support larger, more complex operations. — Lenley Hensarling, technical advisor, Aerospike

Enterprises Will Augment GenAI with Real-Time Data

The true value of GenAI is realized when integrated into enterprise applications at scale. While enterprises have been cautious with trial deployments, 2025 will be a turning point as they begin to scale GenAI across critical systems like customer support, supply chain, manufacturing, and finance. This will require tools to manage data and track GenAI models, ensuring visibility into data usage. GenAI must be supplemented with specific real-time data, such as vectors and graphs, to maximize effectiveness. In 2025, leading vendors will begin rolling out applications that leverage these advancements. — Lenley Hensarling, technical advisor, Aerospike

Increased Skepticism Toward AI Content

AI-generated images and text have been under fire for the past year and I cannot see that changing any time soon. Some consumers report AI generated imagery to seem "cheap" and like "cutting corners." Without any major advancements in the field, it's likely that the trend will continue. — Karolis Toleikis, CEO, IPRoyal

Generative AI Slowdown and Specialization

GenAI has been booming for the past couple of years, however, we are already seeing slight slowdowns in the industry. I think we'll see that continuing while the industry works its way towards highly specific applications in business. Additional, highly specific models may also be developed. AI models are running out of data and are attempting to invent new ways to continue improving LLMs while bypassing the data-hungriness. We'll also likely see greater rigor requirements for AI models as both governments and businesses start implementing them more broadly. — Karolis Toleikis, CEO, IPRoyal

The Rise of AI Factories — Powering the Next Era of Generative AI

In 2025, AI factories will drive a new wave of technological growth and play a pivotal role in transforming how businesses operate and compete. We anticipate double-digit growth in these specialized facilities over the next 3-5 years, with a focus on sovereign AI data centers that prioritize privacy and security. With this, communications service providers (CSPs) are uniquely positioned to complement hyperscalers by leveraging their data centers, existing experience in monetization, and deep relationships with enterprise and government customers. With their robust network infrastructure, CSPs can power AI-generated content while bundling it with connectivity services, presenting new opportunities to finally monetize 5G/6G technologies. A strong platform in 2025 will be key to unlocking revenue potential by optimizing infrastructure and enabling enterprise-grade models and generative AI applications as a service across industries. — Doron Sterlicht, head of R&D, Amdocs

Agentic AI Is Coming for Retail

eCommerce providers will start to experiment with agentic AI in 2025, but the technology won't take off for another 3-5 years. Agentic AI will be applied to inventory optimization and help with stock management across different fulfillment sites. The technology will also benefit customer communication by relaying personalized, timely updates on fulfillment status and timelines. — Steve Sermarini, senior director of data and advanced analytics, Radial

It's OK to Be a Second Mover

In the rush to adopt AI, being a calculated second mover can pay off. Many companies are scrambling to develop half-baked AI-powered products, often without the care usually given to such important product decisions. Motivated by the desire to simply "check the AI box" for customers or investors, these solutions often deliver minimal value at great cost, something not lost on customers. By focusing instead on identifying the most valuable AI use cases for customers' specific needs and doing what it takes to build those well, disciplined companies are learning from early adopters and releasing superior products. Even if they aren't always the first to market, these superior products, and the companies building them, will win. — Michael Zuercher, CEO, Prismatic

AI's Big Price Tag Will Stunt Widespread Adoption in 2025

The economic realities of AI will begin to come to light in the year ahead. While everyone's rushing to integrate AI into their development stacks, we'll soon see companies diving headfirst into AI-powered tools only to find themselves drowning in unexpected costs. The expenses associated with training models, maintaining AI infrastructure, and licensing third-party AI services will create a significant barrier to entry for many organizations and may even stunt widespread adoption in the short term. This economic reality will force a reevaluation of AI's role in software development. It's not about whether AI can do the job — it's about whether companies can afford to let it. The winners in this space won't be those who adopt AI haphazardly, but those who strategically implement it where the benefits truly outweigh the costs. — Tanner Burson, VP of engineering, Prismatic

GenAI/ML Will Enhance Identity Governance  

GenAI/ML are likely to play a more significant role in Identity Governance and Administration (IGA) by simplifying tasks like access requests and approvals, where they can provide valuable guidance and support. However, the effectiveness of GenAI/ML in the deepest aspects of IGA, such as business logic analysis and role mining, may be limited due to bad data hygiene. This often results from inconsistent governance and could skew GenAI/ML insights. Nonetheless, AI/ML will be useful at a higher level, potentially aligning regulatory requirements, business processes and job-related permissions more effectively. The goal of this particular innovation is can a user and chat assistant accomplish their goal with a time and cost reduction? We estimate that the cost per transaction of a user and AI chat assistant will be a fraction of a help desk call price. — Theis Nilsson, vice president global advisory practice, Omada

AI-Driven Transformation in Enterprise Systems

Enterprise systems, including ERP and CRM platforms, will see significant AI-driven upgrades, integrating intelligent agents that enable extensive automation of business processes. These AI enhancements will streamline workflows, reduce manual effort, and empower teams to focus on higher-value tasks.

AI Ethics and Governance: The rapid evolution of generative AI (GenAI) brings immense potential, but it also necessitates stronger restrictions and governance frameworks. It is necessary to ensure these technologies do not fall into the hands of malicious actors or be used in ways that could compromise ethical standards.

Hallucination and Grounding: One of the significant barriers to the broader adoption of GenAI applications is the reliability of their outputs. By emphasizing grounding, we can enhance the reliability and trustworthiness of AI-driven applications.

Explainability in Decision-Making: In areas where AI influences critical decision-making, such as healthcare, finance, or public policy, explainability will be critical to driving reasoning.— Ram Palaniappan, CTO, TEKsystems Global Services

As AI Workloads Become Increasingly Critical, Data Centers Will Undergo Upgrades as AI Workloads Become Increasingly Critical

As AI workloads continue to grow in importance, data centers will go through significant upgrades to accommodate the increased demands of these advanced technologies. This will include boosting power consumption, implementing liquid cooling systems to manage the heat generated by AI hardware, and reconfiguring racks to support specialized AI infrastructure.

Colocation Data Centers: In 2025, we will see a growing trend of AI workloads running alongside traditional enterprise systems in colocation data centers. This hybrid approach will allow businesses to take advantage of AI capabilities while maintaining the security and control of their enterprise infrastructure.

Life of Technology: The rapid evolution of AI technology, particularly with the release of newer AI GPUs every six months, will accelerate the depreciation of existing hardware. This fast-paced obsolescence will put pressure on cloud providers to either write off older equipment or pass on the increased costs to their customers, raising the price of AI services.— Ram Palaniappan, CTO, TEKsystems Global Services

AI and Power Infrastructure Investment and Expansion Remains a National Priority

If AI compute scaling maintains its current growth trajectory, GPU clusters will surge in size from 100K+ to 1M+ clusters, reaching gigawatt scale before 2030. The U.S. is currently the leader in AI globally, maintaining a computing, chip, and technology advantage over its nearest rival, China, with the U.S. having approximately 2X the number of installed computing servers as China. However, since 2000 China has outpaced the U.S. in terms of adding power infrastructure (adding 925 GW of generation via the U.S. increase of 51 GW) primarily in support of its manufacturing base but readily able to pivot to support data center infrastructure.  For the U.S. to maintain its advantage, power infrastructure investment needs to materially expand to the 100+ gigawatt range. Thankfully, this appears to have become a bipartisan area of political concern, and I believe the prioritization around national economic and security interests will help accelerate infrastructure development. However, a question remains, will investment be fast enough to maintain the U.S. technological advantage or will innovation be bottlenecked due to capital or regulatory constraints? — Tom Traugott, SVP of strategy, EdgeCore Digital Infrastructure

Gap Between AI's Promise and Reality Will Widen Without Better Integration

AI's progress will continue, but unless we bridge the gap between promise and reality through workflows, most companies won't see the returns they're hoping for. In 2024, we've seen the "delta" between AI's promise and the reality of putting it to work in an enterprise. Foundational models are getting smarter all the time, but they're only as good as the data they can access. In order to see this gap close in 2025, they should be trained on proprietary business data—the real goldmine that makes a company successful. This is what keeps AI from fully realizing its potential in enterprise settings. The true opportunity, then, is in using workflows to close this gap. The gap between AI's promise and its real-world value lies in seamless integration with workflows. That's exactly what we're seeing with clients who use workflows to connect AI to their systems. The future of enterprise AI isn't just about making AI smarter, but about making it relevant through integration. — Eoin Hinchy, CEO/co-founder, Tines

Transparency Will Be the Key to Building Trust in AI

Transparent workflows will be essential to making AI trustworthy over the next year, allowing people to look "under the hood" and see how decisions are made. When it comes to AI trust, transparency is absolutely essential. If users can't see how an AI solution came to a certain decision, they're going to be skeptical about letting it into critical parts of the business. That's why workflows have a huge role to play in giving users a transparent view of each step of the process. If you ask, "What's our annual recurring revenue (ARR)?" and the AI spits out a number, workflows should let you dig into how that number was arrived at. You'd be able to see which workflow ran, the query made in Salesforce, and the raw results that came back. Transparency builds trust, especially in complex environments. For companies investing in AI in 2025, it's this transparency that makes all the difference between a tool that's useful and one that's just a black box. — Eoin Hinchy, CEO/co-founder, Tines

Push for Measurable ROI on AI Investments Will Intensify

In 2025, it will no longer be enough to just "adopt AI" — companies will need hard ROI metrics to prove its value. We're now a couple of years into the generative AI boom, and I think it's fair to say that the technology hasn't yet lived up to its hype. CIOs and CTOs will demand concrete metrics before approving new AI investments. Going forward, companies are going to need hard ROI to justify spending on AI tools. Metrics like "80% of code now touches AI" or "50% of customer queries are resolved by AI" are going to be essential. It's no longer enough to just demo an AI solution and assume it will add value. We need quantifiable outcomes. And the companies that can show hard data on cost savings or productivity gains are the ones that will actually see AI succeed in their business. — Eoin Hinchy, CEO/co-founder, Tines

AI Revolutionizes Data Classification

Data classification is one of the first significant data security problems AI can effectively solve. The ability of AI to accurately classify vast amounts of data will help organizations better manage sensitive information, reduce false positives and negatives, and improve overall data security posture. This advancement will be crucial as data volumes and complexity continue to grow. AI-driven classification systems will become sophisticated enough to understand context and intent, not just content, leading to more nuanced and accurate data protection measures especially for challenging unstructured data sources. And we'll see longstanding data governance and compliance challenges solved, enabling organizations to automate many previously manual and error-prone data protection aspects. — Ron Reiter, CTO/co-founder, Sentra

Enterprises Will Unlock AI Potential in 2025 Through Cost Optimization, Empowered by FinOps

I believe enterprises will transform AI project viability through more purposeful cost optimization, with FinOps teams becoming critical enablers. I expect we'll see many more organizations begin to pair AI initiatives with automated provisioning, real-time cost data, and sophisticated cost controls to reduce their AI investment risks. Getting this right will turn potentially unsustainable long-term AI initiatives into measurable, financially responsible innovation engines. — Tzvika Zaiffer, director, Solutions, Spot by NetApp

AI Will Become Indispensable Business Advisor

AI will cement its position in 2025 as an indispensable business advisor, moving decisively beyond experimental status. By providing nuanced risk analysis, surfacing hidden opportunities, and delivering contextual research with real-time data access, AI will become very deeply integrated into enterprise decision-making. This evolution marks a crucial stepping stone toward AI's future role as an autonomous decision-making agent. — Anil Inamdar, head of consulting services, NetApp Instaclustr

Data Quality Supersedes Quantity, Placing a Greater Onus on AI Customers

We're seeing growing reports that LLM providers are struggling with model slowdown, and AI's scaling law is increasingly being questioned. As this trend continues, it will become accepted knowledge in 2025 that the key to developing, training and fine-tuning more effective AI models is no longer more data but better data. In particular, high-quality contextual data that aligns with a model's intended use case will be key. Beyond just the model developers, this trend will place a greater onus on the end customers who possess most of this data to modernize their data management architectures for today's AI requirements so they can effectively fine-tune models and fuel RAG workloads. — Rajan Goyal, CEO & co-founder, DataPelago  

AI Factories Evolve to PaaS

In 2025, AI factories will evolve beyond their initial phase of providing infrastructure-as-a-service, offering compute, networking, and storage services, to delivering platform-as-a-service capabilities. While the foundational services have been essential to jumpstart AI adoption, the next wave of AI factories must prioritize platforms that drive data affinity and provide lasting value. This shift will be key to making AI factories sustainable and competitive in the long term. — Rajan Goyal, CEO & co-founder, DataPelago  

Agentic AI Poised to Revolutionize Workflows

Agentic AI will be everywhere — and it's going to make our lives easier agentic AI will take on tasks that will significantly alleviate the burden on human time. It will be able to create AI team members who take on tasks and can perform actions, such as carrying out research, summarizing and aggregating their findings (and critiquing and improving those before presenting them to a human), creating reports and plans, composing and sending emails on a human colleague's behalf, and more: in a fraction of the time it would take a human. All of this means AI will start using software applications itself rather than just being behind the scenes. In turn, this will need re-evaluating the traditional business models used to charge for software. — Rod Cope, CTO, Perforce

Building Trust into AI Becomes Critical

Especially within regulated industries, we will see more effort going into compliance, governance, auditing, transparency, and explainability: In other words, all the elements that will help make AI more trustworthy. Not only is it crucial to have this trust built into AI in terms of privacy, security, and compliance, but creating more trust in AI will also improve confidence around its use, especially by organizations that, at the moment, are holding back. Proper governance around AI will be a top priority in 2025. In addition to tools from organizations such as Perforce, we can expect to see other tools introduced, especially those that use AI to "police" AI. — Rod Cope, CTO, Perforce

AI-Enabled Tech Is Going to Help Us Humans Keep Up with AI — and So Much More

The concept of bidirectional brain-machine interfaces is a particularly interesting one. These interfaces can read people's thoughts and communicate them outwards-and also vice versa, such as affecting thoughts, emotions, and memories. Of course, both these devices (plus robots) raise huge questions around ethical use, but putting those aside for one moment, bi-directional brain-machine interfaces could help humans keep up with AI by enabling them to think faster. For instance, recent breakthroughs will use AI to reduce drug development and delivery from many years to just months. Imagine if you could then add AI-enhanced humans to the equation, and discoveries and decisions could be made even faster. — Rod Cope, CTO, Perforce

Those Sci-Fi-Style Predictions Could Be Sooner Than Many of Us Expect

Gartner predicts that by 2030, 80% of humans will engage with physical robots on a daily basis. They are expected to handle manual tasks, especially where there is a labor shortage and in some aspects of healthcare (for instance, to address the nursing shortage). Likewise, Gartner predicts that 30% of knowledge workers will be enhanced by and dependent on technologies such as bidirectional brain-machine interfaces by 2030. — Rod Cope, CTO, Perforce

2025 Marks the 'AI Pivot' as CFOs Demand Measurable ROI from AI Investments

In 2025, the "AI Pivot" will take center stage. Over the past 18 months, the industry definitely assigned inflated expectations to AI, thinking it could do everything. Simply mentioning "AI" once seemed to promise miraculous results. However, in the last six months, CFOs have begun to push back, questioning the ROI associated with substantial investments in AI technology. In the year ahead, businesses will focus more on balancing the use of AI with delivering measurable business results. Companies will need to bridge the gap between expanding AI capabilities and ensuring these investments drive revenue growth and/or cost reduction. The initial hype around AI will give way to more deliberate and scrutinized investments, with an emphasis on practical applications that boost employee productivity, reduce costs, and improve IT service management. Included in the ability to thrive in this evolving AI landscape will be the need for companies to prioritize how to leverage AI to strengthen security measures while remaining cautious of its potential security risks. — Doug King, CIO, ePlus

State-Level AI Legislation Will Ignite a New Wave of AI Legislation and Test American AI Leadership

California and Texas are poised to lead a transformative era of AI regulation, setting the pace for other states with legislation addressing urgent challenges like ransomware, LLM safety and oversight, and ethical AI use. However, state-specific rules may create friction with federal policies and complicate compliance for businesses operating across state lines, increasing costs, added compliance, and operational hurdles to navigate a state network of patchwork legislation. The lessons of past state privacy legislation and federal inaction may be a comparable experience. As the patchwork of state laws grows, pressure on the federal government to act will intensify. A unified approach will be critical to minimize economic impacts and ensure innovation is not stifled. An outstanding question is whether the new Republican-controlled Congress can prioritize with the Trump Administration on rules of the road in a manner that can keep the United States ahead of its AI race with the Government of China. Concerns over Chinese AI advancements may create bipartisan cooperation, and establish potentially unlikely alliances, but the question is how quickly Congress can legislate when it is likely that the Trump Administration will revoke the current Biden White House AI Executive Order, which has worked in parallel with the Senate's AI process, led by Senator Schumer (D-NY) and Senator Rounds (R-SD). While these federal regulations could create compliance challenges, they may also offer new opportunities by fostering a safer, more ethical AI landscape if it can satisfy fears of losing pace with Chinese innovation. — Jeff Le, VP of Global Government Affairs and Public Policy, SecurityScorecard

The Rise of Agentic AI Will Require a Rethinking of Security Strategy

Generative AI is quickly moving beyond the capabilities of consumer-first tools like ChatGPT into agentic AI for the enterprise. AI agents are designed to process information in a new way to make dynamic and autonomous decisions. However, organizations looking to leverage the promise of agentic AI need to be wary of the security ramifications. They can do so by going beyond analyzing prompts and responses by monitoring and profiling how each AI Agent operates behind the scenes. Given the widespread access these Agents have to sensitive information, this holistic approach can prevent direct and indirect prompt injection attacks, as well as help to manage data leakage risks. Staying secure amid new threats will require security teams to work with the business not as a blocker but as an enabler. — Ben Kliger, CEO and co-founder, Zenity

A Culture of Innovation Will Be the Catalyst for AI Success

AI success requires more than just implementing new technology — it demands embedding innovation into an organization's cultural DNA while streamlining their technical foundation. Looking ahead, companies will need to empower team members across all levels to prototype and deploy AI solutions rapidly. Some organizations already require every intern to build and submit AI projects. This cultural shift around AI-powered transformation must be combined with actively reducing technical complexity, creating an environment where experimentation and iteration are encouraged and expected. Companies that can rapidly prototype, measure results and scale initiatives through integrated systems will thrive in their markets. Organizations that embrace swift experimentation over perfect implementation will build resilient AI capabilities that evolve with market demands. — Rich Waldron, CEO and co-founder, Tray.ai

AI Progress in Data-Rich Industries Soars

We will continue to see great advancements for AI in industries where data is plentiful, such as healthcare, but we'll see brands struggling to activate AI in meaningful ways for their consumers. Most of this will be driven by the discovery that most of their data is unstructured, incomplete, and is full of biases due to how digital data has been captured over time on their websites and apps. We will see a rise in stories of poor uses of AI as a result as well, which will cause brands to pump the brakes a bit and revisit their data strategies. — Bill Bruno, CEO, Celebrus

Generative AI Moves into Everyday Tools

Generative AI adoption will increasingly be embedded in everyday tools, such as conferencing software, Microsoft applications, and GitHub Copilot, enhancing user experience and productivity. As businesses move beyond experimentation, AI will enter production environments, enabled by advancements in governance, privacy protections, and solutions to talent shortages. — Michael Curry, president of data modernization, Rocket Software

AI Premium Pricing Model Will Eventually Collapse as Features Become Table Stakes

The current model of charging premium prices for AI features as add-ons will face increasing pushback from enterprise customers in 2025 and beyond. With AI becoming a standard in tech stacks, AI processing must become more cost-efficient. This shift and customer expectations that AI should enhance offerings and not raise software costs will force vendors to make AI capabilities a standard and integrate them into core product pricing. — Julie Irish, SVP and CIO, Couchbase

AI Success Rates Will Improve as Organizations Test Smaller Pilot Programs

The AI hype cycle is unlikely to ever truly die. However, we're seeing a shift in AI perception as more leaders acknowledge the inherent limitations of AI-driven technologies. In this new era, businesses will shift their focus from chasing the latest AI buzzword or trend to solving tangible problems. Too often, companies jump into AI without a clear strategy, asking, "How can we use AI?" rather than, "What business problems actually need solving, and can AI be a part of the solution?" In 2025, organizations that thrive will prioritize aligning AI with specific goals, such as automating repetitive processes, improving customer service, or optimizing resource allocation. This shift requires identifying high-value, low-effort projects to generate early wins and build organizational confidence. For example, automating customer call routing can deliver measurable ROI quickly, setting the stage for larger initiatives. By treating AI as a tool for solving business challenges rather than as a magic solution, industry leaders will see more successful AI programs. — Mike Simms, vice president — Data & AI, Columbus Global

Agentic AI Poised to Power Next Wave of AI Innovation in 2025

In 2025, the trajectory of AI will be shaped by the rise of agentic AI — proactive, intelligent agents that go beyond basic chatbots in an evolution promising a profound transformation for both consumer and enterprise landscapes, accelerating the world into a new era of AI capabilities. With capabilities such as understanding context, setting goals and adapting actions, agentic AI can complete tasks previously thought impossible by AI. For this to be made possible, agentic systems require a compound AI system using multiple models that are moved closer to data sources, within security parameters. The systems also need to handle both structured and unstructured data at low latency — all in real-time — to make meaningful, context-aware decisions on the fly. This requires seamless integrations across unstructured data processing, vector databases and transactional systems for efficient storage and retrieval of diverse data types. The companies that will excel in providing these robust integrations and infrastructures will be uniquely positioned to drive the next wave of innovation and value in the AI sector. — Rahul Pradhan, VP of product and strategy, Couchbase

Edge AI and Vision Combine to Revolutionize Manufacturing and Logistics

In 2025, I expect we will see a push toward the use of edge AI and vision in combination to drive automation, especially for manufacturing and logistics industries. Processing vision data at the extreme edge, on sensor, will allow warehouses and factories to be agile with their cloud costs, sending only specifically trained metadata responses to the cloud, instead of costly masses of image and video data. As these industries are pushed to transform, needing to make the most of strained time and financial resources, where many employees may have little or no previous AI development or engineering experience. But with technologies like edge and vision AI for resource-limited warehouses and employees can benefit from automated defect detection, safety checks, berth efficiency, and beyond. — Eita Yanagisawa, senior GM of System Solutions Business Division, Sony Semiconductor Solutions, AITRIOS

AI Set to Revolutionize Fashion in 2025

Artificial intelligence is rapidly transforming the fashion landscape. Algorithms make it possible to anticipate trends, optimize inventories and even create designs. At the same time, AI is personalizing the customer experience to unprecedented levels, offering a competitive edge in a saturated market. In 2025, the most technologically agile brands will leverage these innovations to accelerate their growth while adapting to a frantic pace. — Lenny Marano, president Americas, Lectra

Human-AI Collaboration in 2025: Balancing Creativity and Automation

The dynamics of human-AI collaboration will evolve, enabling humans to prioritize strategy and creativity. Generative AI and machine learning will enhance content efficiency and automation while also raising concerns around quality and misinformation, driving demand for transparent AI use. Brands will have the opportunity to balance automation with human oversight, leveraging solutions like Knowledge Management and Retrieval Augmented Generation (RAG) that rely on quality human inputs. Preparing content for AI, using consistent metadata and knowledge graphs will be essential to obtain reliable AI outputs and develop specialized LLMs tailored to their needs. — Thomas Labarthe, president of Content Technologies, RWS

AI Value in Business

Today, businesses do not have an intuitive understanding of where AI value lies. They have an understanding of where their processes are inefficient, and where customer needs exist, but are still attempting to develop an understanding of how to apply AI capabilities to these problems. The progress here will be more of a matter of incremental progress as the market responds, as opposed to an exponential takeoff in the general case. — Zachary Hanif, VP of data and AI/ML, Twilio

Expect a Grassroots Acceptance of AI Output

Trust and authenticity are going to become increasingly important to discerning participants in the market for generative experiences, but for human assistance, and editorial functions, this effect will not be as pronounced. Instead, we will observe a more grassroots acceptance of AI output with resistance from existing structures. — Zachary Hanif, VP of data and AI/ML, Twilio

More Companies Will Run Customized AI Models On-Premises

In 2025, we will see a shift toward on-premises AI deployments. As open-source models become more cost-effective and accessible, organizations will increasingly opt to run customized versions within their data centers. As a result, it will be cheaper, faster, and easier to own AI models and fine-tune them to individual needs. Companies will find they can combine their data with existing models and tailor the experience for their customers at a fraction of today's costs. Meanwhile, increased compliance risks associated with AI will drive regulated industries, like financial institutions and government agencies, to deploy models in air-gapped environments for greater control over data privacy and security and reduced latency. — Emilio Salvador, VP of strategy and developer relations, GitLab

AI Agents Will Be Catalysts for Software Supply Chain Transformation

AI agents are poised to revolutionize the software supply chain by automating and optimizing processes, from continuous integration to continuous deployment. This transformative shift will initially gain traction in open-source ecosystems, where AI agents will likely be built and shared with the community, like software libraries. As developers and organizations witness the benefits of AI-driven automation in open-source projects, we can expect a rapid expansion into commercial enterprise solutions. Internal development teams and platform engineers will increasingly be tasked with building, extending, and integrating AI agents across the entire software supply chain. — Lee Faus, global field CTO, GitLab

AI Will Drive Efficiencies for Platform Engineers

The proliferation of pattern recognition in AI technologies is expected to reduce the friction of automating software releases into production. By creating reusable building blocks that encapsulate common functionalities for software delivery, platform engineers will help empower non-technical team members to easily assemble delivery pipelines using intuitive low-code techniques for testing, environment management, and release orchestration. This movement will lead to a rise in application development driven by AI-assisted tools, enabling organizations to meet specific needs more efficiently. — Lee Faus, global field CTO, GitLab

CIOs Need to Prepare for Agentic AI to Flip Workplaces on Its Head

As businesses integrate AI into everyday processes, organizations must prioritize communication and reskilling their workforce now, and continue education throughout 2025. CIOs know technologies like AI agents are poised to change the workplace, but they need to get ahead of workers' fears that it is coming in to replace them. AI's role is to augment their jobs, not take them. Businesses that fail to proactively address employee concerns around introducing agentic AI especially, risk resistance and inefficiency in implementing these technologies. We've seen the data and it's clear: Early adopters of GenAI are the current winners, their employees are the winners, and the early adopters of AI agents are sure to follow a similar course. — Carter Busse, CIO, Workato

AI Initiatives Will Be as 'Unsuccessful' as Business Leaders Make Them

A question every business leader is asking is whether or not the investments they have made in AI have produced anything of value. They are asking this question too soon. In 2024, many organizations threw money at AI with the mindset that they would see immediate and meaningful results, without thinking critically about what those key indicators are.  Now that we have experimented with AI in 2024, this year we will see leaders determine the key metrics for evaluating success and thinking about long-term measurement. — Carter Busse, CIO, Workato

Role-Based Agents Could Suffer Same Fate as SaaS: Messy Sprawl

We're starting to see AI agents come up more and more and it seems like every business leader is now thinking about what their roadmap is for AI agents. This is smart because I believe this year we will see the most significant impact on business operations with agents that are role-based. This means workers in customer service or IT for example will become more proficient in their role because agents will become stronger and more capable of handling niche tasks much faster than a person can on their own. In turn, this will allow workers to spend more time learning tasks that require deeper thinking and, in turn, become more knowledgeable in their role. What I see becoming a challenge is the proliferation of AI agents that can handle all these individual tasks will create a web of specialized use cases that don't work together, and overwhelm business users. Similar to what we've seen with the explosion of SaaS applications over the last decade, which companies are still struggling to untangle. — Bhaskar Roy, chief of AI Products and Solutions, Workato

AI Startup Profit Bloat: Experimental Revenue Does Not Translate to ARR

AI startups benefited significantly in 2024 from the budgets companies put towards experimenting with AI applications, generating revenue quickly and grabbing the attention of VC investors. We can't look at this experimental recurring revenue as ARR because a year from now businesses will shift from experimenting with AI to putting it into production — and they'll be keen to realize their investments. We are going to see some of the AI startups lose momentum in the market as their growth turns to churn if they don't look at their books closely. They need to understand now what is truly sustaining revenue and what's not, and adjust their business model accordingly. — Bhaskar Roy, chief of AI roducts and Solutions, Workato

Early-Stage LLM Vendors Will Be Vaporized by the Incumbents

A lot of people in the AI sector questioned in 2024 who will emerge as the winning LLM vendor. I don't foresee there being one top dog, but I believe the incumbents that have dominated market share will either vaporize many startups or acquire them. While startup LLM vendors have the agility and can build things faster, what they lack is the inherent trust with customers that the big players have already established from a security and governance standpoint. With that, customers will be faced with making a leap of faith and investing in a startup, which comes with risk, or doubling down on what they know works. Not to mention, as AI matures we're starting to see the cost of things like compute and AI tokens drop, and that will just keep projecting downward for the established LLM vendors who have a large amount of funding to innovate in ways startups don't. — Bhaskar Roy, chief of AI Products and Solutions, Workato

AI Will Expose Architecture's Breaking Points

While AI can make code development simpler and faster, it won't solve architectural challenges. When it comes to building, improving, and fixing applications, AI tools like ChatGPT and other large language models cannot effectively address issues in the interactions between components. AI code generators excel at writing individual components but miss the bigger picture of how systems work together. They can't grasp how multiple systems interact in real-world scenarios, leaving teams ill-equipped to solve scalability and reliability problems stemming from poor architecture rather than code quality. As AI speeds up coding, teams must focus more on system-wide architecture design and documentation. Understanding architecture, identifying sources of technical debt and complexity, and documenting systems will become increasingly important as AI handles more code generation and makes it easier to churn out software. — Moti Rafalin, CEO and co-founder, vFunction

Generative AI Will Redefine the Boardroom

In 2025, generative AI will reshape business strategy. Today, 99% of enterprises are integrating AI into their revenue processes — but the next leap is transformative. Picture AI models delivering real-time recommendations to navigate complex markets, optimize revenue flows, or counter economic headwinds. Boardrooms will evolve from static reports to interactive, AI-powered solutions that simulate future scenarios with unmatched precision. Decisions will no longer rely on hindsight — they'll be driven by AI's ability to chart the smartest, most strategic paths forward. — Andy Byrne, CEO, Clari

The Age of Autonomous Business: How AI Will Drive Revenue Growth

AI will evolve from an assistive tool to the operational backbone of business growth. Enterprises will deploy autonomous systems that dynamically manage decisions, optimize workflows, and eliminate inefficiencies. Entire industries will adopt "self-driving" revenue systems that predict outcomes and take action, enabling leaders to focus on strategy instead of execution. — Andy Byrne, CEO, Clari

AI in Customer Service

After years of speculation and anticipated hype, 2025 will truly be a turning point for artificial intelligence in business operations — this will be the year it delivers tangible business value, particularly in customer service operations. Call centers will be at the forefront of this transformation, with AI assistants becoming not only equivalent to but exceeding human performance in handling things like technical support, billing questions, and password resets. Chatbots will become increasingly indistinguishable from human support agents, and AI will evolve from its current "assistant" role to handling more specific roles independently. — Russ P. Reeder, CEO, ATSG

Robotics Will Transform Daily Life and Work

In 2025, robotics is set to enter a new era of accessibility and integration. We're on the brink of a time when robots aren't just for specialized industries — they will become part of our daily lives, helping to streamline everything from healthcare to construction. As robots become more autonomous, they'll take on increasingly complex tasks, moving beyond simple, repetitive actions to more adaptive and dynamic rules. This shift is driven by AI and advances in the creation of simulation environments to train them, which enables robots to learn how to make intelligent and independent decisions and work together across industries in ways that have never been seen before. Additionally, in the next few years, we will see smart, collaborative robots integrated into sectors like manufacturing, energy and agriculture. The future of robotics is already unfolding and it is transforming how we work and live. — Agustín Huerta, SVP of digital innovation, Globant

AI Governance Takes Center Stage

2025 will be the year of getting AI under control — or, in other words, AI governance. Now that organizations know the value of getting AI into production, they realize that controlling cost, quality, and access is critical for making the most of the technology. People will also be reminded over and over again of the consequences of the fact that AI doesn't know what it's talking about — or as Stefan Wrobel put it "AI states the likely, not the truth — but does so remarkably well." For some applications this is fine, but for most, this is a fundamental problem. How to make AI reliable will be a key focus of 2025. And finally, 2025 will be the year AI is overused. As the expression goes, to someone who holds a hammer, everything looks like a nail. But after everyone better understands the limitations and costs of AI, I see people often returning to classic analytics and text analysis methods that are cheaper, easier to control, and more reliable. The most powerful and innovative approaches will combine those classics with new techniques. — Michael Berthold, CEO, KNIME

AI Breakthroughs and the Race to Exceed OpenAI's o1 Model

2025 will be a breakthrough year for AI milestones, as competitors race to keep up and exceed with OpenAI's o1 model. But these advancements in LLMs ability to "reason" will be tempered by acknowledgement of real concerns, including how more powerful AI models require increased computing resources. 2025 will be the year AI leaders confront balancing innovative capabilities with cost and energy limitations, as well as customer use cases. —Robert (Bobby) Blumofe, CTO, Akamai

The Rise of AI-Powered 'Supervisors'

Over the next year, we will witness the evolution of enterprise AI agents as they become increasingly sophisticated in their reasoning and comprehension capabilities. Emerging use cases will transform the way businesses leverage these agents, and the nature of human interaction with them will evolve as they take on more complex tasks and decision-making roles. Watch for the rise of AI-powered "supervisors," that will have the ability to move past simply automating tasks to truly orchestrating the interaction of all AI agents throughout an entire organization. This will make it exponentially easier for humans to administrate teams of AI agents across their entire business. By the end of 2025, AI agents will cross the chasm from tools that require more hands-on supervision to fully autonomous systems. Expect to see AI agents independently automating complex, multi-step processes without a human in the loop. This will undoubtedly transform how executives view AI adoption, positioning it as a powerful engine for unprecedented growth and innovation. — Dorit Zilbershot, VP of AI and innovation, ServiceNow

Industry-Specific GenAI Demand Takes Off

Over the next year, demand for industry-specific GenAI — context-aware and bespoke to specific industry use cases — will grow exponentially, as the technology continues to mature and offer highly specialized and impactful solutions. This shift could upend traditional market dynamics over the next several years, sparking growth in verticals like telecom (e.g., rural broadband), finance (e.g., regulatory tech) and healthcare (e.g., telemedicine for rare diseases). These once lower margin sub-sectors and services that historically struggled to break through due to high operational costs, limited customer bases, or less efficient business models, will attract new investment as GenAI allows them to entirely reinvent business models to operate more efficiently and profitably. — Dorit Zilbershot, VP of AI and innovation, ServiceNow

Embracing a Hybrid AI Future

In the field of AI, predicting the future is about as reliable as a weather forecast, especially in the era of large language models (LLMs). The truth is, we're headed toward a hybrid future. It's important to note that both open and closed-source models have their place, despite the popular sentiment of open-source takeover. Enterprises are better off being model-agnostic. The open-source vs. proprietary discourse does no good to an organization building robust solutions capturing the best of both worlds. Closed source models, developed by well-resourced companies, often push the boundaries of what's possible in AI. They can provide highly refined, specialized solutions that benefit from significant investment in research and development. — Sreekanth Menon, global head of AI, Genpact

Industries Will Shift Focus to Agentic AI

Without a doubt the top technology that industries will be focusing on is getting to the next stage of GenAI using agentic AI. While previous changes were about optimization and efficiency, the stage is now set for having a multitude of expert AI agents feeding data from industry specific knowledge and data stores, filling the gap between the general knowledge in LLMs and the plethora of data generated in the last decade. — Brendan Bonner, innovation lead, Office of the CTO, Extreme Networks

Companies Identify Top-Performing Ideas to Drive AI/ML Investments

AI will gain more clarity in 2025 as some of the POCs run their course and companies find the highest 1-2 productive ideas to drive their AI/ML investments more clearly. — Jim Kozlowski, chief sustainability officer and VP Data Center Operations, Ensono

AI Will Revolutionize Meeting Spaces

As people return to office and hybrid work becomes more common, AI will play a key role in transforming how we interact with meeting spaces. Instead of a rigid, one-size-fits-all approach, AI will learn your preferences — adjusting the lighting, displays and the meeting platform to your liking as soon as you enter a room. Whether you're using Zoom, Teams or another tool, AI will ensure the space is automatically configured to match your needs, saving you the time and hassle of manual adjustments. The goal is for AI to recognize your credentials and preferences, regardless of the platform or service, so you can walk into any space and be ready to go. Ultimately, this will create a seamless, consistent experience that enhances collaboration. —Dan Root, head of Global Strategic Alliances, Barco ClickShare

The Financial Reality & Risks of Autonomous AI Agents

While AI agents offer remarkable promise for boosting productivity and automating security, they also come with significant operational costs and risks. Running these autonomous agents in a live development environment requires ongoing, resource-intensive model calls to analyze, propose, and validate code changes, which can quickly drive up expenses. For organizations planning large-scale deployments, balancing the cost of AI implementation with productivity gains will be essential. Additionally, security remains a pressing concern; AI agents are vulnerable to prompt injection attacks, where adversaries can manipulate them into unintended actions. This limitation underscores the need for continuous human oversight. In 2025, organizations will need careful strategies to ensure AI agents are both economically viable and securely managed to truly benefit from their potential. — Randall Degges, head of developer and security relations, Snyk

AI Contributing to Autonomous Industrial Operations

AI combined with traditional deterministic automation and sensing technologies is providing a foundation towards increasingly autonomous industrial plant operation — improving safety, reliability, and efficiency in facilities and removing people from hazardous locations. — Jason Urso, VP and CTO of Industrial Automation, Honeywell

AI Moves from Flashy Experiments to Solving Real World Problems

The next phase of AI evolves from unstructured and unpredictable LLMs looking for problems to solve to more process simplifying use of agents to improve productivity and decision making. — Ross Meyercord, CEO, Propel Software

Applications of Agentic AI

Software that can reason — that can plan and execute steps to get something done on behalf of the organization. This can effectively flip the role of software on its head, so that intelligence running in the cloud carries out tasks for you. There are both employee-facing and customer-facing applications: essentially, in any area where processes need to be streamlined or made more effective. Improving customer experience and satisfaction; an agent can come in and solve more of the problem for the customer, without manual intervention — and probably with fewer errors — or escalate quickly to a human based on their judgement and sentiment analysis. For example, in order management, especially as manufacturing becomes more automated and it gets easier to create just-in-time custom-ordered products, customers may be able to contact a business more readily with tweaks to their product or to request updates. — Gordon Van Huizen, SVP of Strategy, Mendix

Increased Adoption of AI and ML, Especially ML and RPA

The prevalence of generative AI (Gen AI) will significantly drive automation across various sectors, including small to medium-sized businesses (SMBs), with the goal of enhancing operational efficiency. This trend is unfolding due to the rapid advancements in AI technologies, which make complex tasks previously requiring human input more manageable and less resource intensive. In fact, IDC is projecting generative AI spending to reach $337 billion by 2025. — Eric Stavola, vice president of managed services sales and programs, Visual Edge IT

AI Agent-Focused Experience

As AI agents become increasingly capable of performing tasks such as controlling systems and interacting with various software platforms, they will introduce significant challenges—and opportunities—in the realm of data protection and data classification. The rise of AI agents with autonomy to perform complex tasks raises new concerns around security, as they operate on sensitive data, interact with external systems, and are involved in critical decision-making processes. The rise of AI agents is also expected to influence the freelance market, with more companies looking to streamline and specialize instead of hiring full-time talent. — Eric Stavola, vice president of managed services sales and programs, Visual Edge IT

The AI Investment Frenzy Comes Down to Earth

2024 was a banner year for AI investment, with 35% of startup dollars going to AI companies compared to 15% in 2021. The party won't end in 2025, but the cover charge will get a lot higher. With less funding being thrown at foundation models, more will target use cases that yield demonstrable value on top of those models that drive rapid growth in annual recurring revenue. Venture capitalists are already becoming more demanding; the median time lapse between Series A and Series B rounds in 2024 was 28 months, the longest in over a decade. While we aren't looking at a dot-com bubble-like collapse just yet, we can assume that business models with solid unit economics will get the lion's share of venture dollars in 2025. — Jeremy Burton, CEO, Observe

The Rise of AI Agents in the Enterprise

Much has been made of AI's potential game-changing impact to consumer search, but perhaps some of the biggest short-term opportunities in AI may lie in addressing the woeful inefficiency of many day-to-day office tasks. ERP and CRM systems may be the backbone of the enterprise, but most of the work still gets done in emails and spreadsheets. Agentic AI promises to do what robotic process automation didn't: scour through that mess, eliminate ROT (redundant, obsolete and trivial) data and make what's left part of actionable workflows. That isn't sexy stuff, but it will give birth to many more successful businesses over the next five years than training ever larger LLMs. — Jeremy Burton, CEO, Observe

Bloom Comes Off the LLM Rose

Despite the excitement around AI, many will begin to realize the limitations of large language models. While impressive at summarizing, translating and regurgitating well-known information, these models are clearly not the foundation for artificial general intelligence (AGI). In fact, LLMs have arguably siphoned investment dollars away from approaches that may have had a greater chance of success. It's not that LLMs are useless — quite the opposite — they are a better way for humans to interact with all kinds of software, devices and systems and will fundamentally change many industries. But the AI super-intelligence — the kind that interacts and reasons about its knowledge, surroundings and people in the same way humans … that's still many, many years away. — Jeremy Burton, CEO, Observe

The Year of Agentic AI

2025 will see the rise of generative AI agents used to solve problems — an approach that is made possible by decreasing costs and increasing the performance and speed of LLM. Frameworks for orchestrating agentic AI work, which refers to the ability of AI agents to act autonomously and make decisions, will emerge, and a large percentage of use cases will begin to employ this approach. Envision one LLM crafting software code while another ensures it's secure, a third checks for style rules, and a fourth optimizes for performance. These agents will iterate multiple times, each iteration bringing them closer to an optimal solution. — Alan Jacobson, chief data and analytics officer, Alteryx

AI Use Will See Improved Cost and Performance

In 2025, we will continue to see order-of-magnitude improvements in the key areas of hardware and algorithms for cost, speed, and accuracy. Some of these gains will be 'eaten up' by the increased use of iterative (agentic) approaches. As the cost and performance gains move, the approaches to solving problems using the 'creativity' of LLMs will be unlocked more fully. — Alan Jacobson, chief data and analytics officer, Alteryx

Models Will Find Their Purpose

Throughout 2023 and 2024, there were significant breakthroughs in the capability of foundational models to perform various tasks better than others. While many companies searched for one ultimate model they would select and use, in 2025, more organizations will realize that different models will be more successful depending on the use case. This diversity may converge in the future, but 2025 will see more fragmentation, with more models emerging as clear leaders for specific use cases based on LLMs' strengths and weaknesses. These will fit within technical categories versus domains or industries and include, for example, rapid summarization of copious amounts of data. Use cases such as highlighting key events that could impact a company, an individual's travel, or a political process would be an example of a genre in which LLMs excel. — Alan Jacobson, chief data and analytics officer, Alteryx

CIOs Will Be Held Accountable When AI Failings Occur

In 2025, as AI innovation and exploration continues, it will be the senior-most IT leader (often a CIO) who is held responsible for any AI shortcomings inside their organization. As new AI companies appear that explore a variety of complex and potentially groundbreaking use cases, some are operating with little structure in place and have outlined loose privacy and security policies. While this enables organizations to innovate and grow faster, it also exposes them to added confidentiality and data security risks. Ultimately, there needs to be a single leader on the hook when AI fails the business. To mitigate potential AI risks, CIOs or IT leaders must work closely on internal AI implementations or trials to understand their impact before any failings or misuse can occur. — Joel Carusone, SVP of data and AI, NinjaOne

AI Funding Crunch Will Accelerate M&A

In 2025, we'll see more consolidation in the AI market. AI organizations require immense resources to sustain innovation and manage infrastructure. It's expensive to be an AI organization today. In order to continue growing at this expedited clip we're seeing, AI companies will be fundraising over the next 18 months. Some will succeed in raising capital. Others will be consumed by larger players via M&A. — Joel Carusone, SVP of data and AI, NinjaOne

More Than Half of Retailers Will Invest in AI Platform Technology

As retailers recognize the value of unified AI solutions over piecemeal approaches, we predict that over half of them will adopt AI platform technology to support a growing range of business applications. This platform approach will enable retailers to apply AI-driven insights across business functions such as loss prevention, inventory management, and customer experience. With the tech industry increasingly focused on this market, retailers are well-positioned to integrate foundational AI with tailored applications. — Alex Siskos, SVP of strategy, Everseen 

Businesses Shift Toward Localized, Privacy-First AI Solutions

Business's AI strategy will continue to pivot — dramatically, for many — in 2025 toward localized, privacy-first AI solutions. There's an accelerating recognition that true digital transformation requires AI assistants that can *securely* interact with proprietary data without exposing sensitive information to external networks. The rise of local AI models will transform how businesses approach generative AI, prioritizing security, compliance, and precise data control over broad but potentially risky cloud-based alternatives. — Brian Sathianathan, CTO and co-founder, Iterate.ai

Addressing Emerging AI Liability Risks Amid Growing Adoption

I do think we should consider the potential liability of AI models themselves as an emerging risk. We as an industry should be considering who is liable if an AI-driven recommendation causes harm. As AI becomes more widely adopted, there will be more of a need to ensure and regulate AI models against unforeseen consequences. — Leandro DalleMule, general manager, Planck / Applied Systems

AI's Future Depends on Hardware Availability, Efficiency

One important AI-related area I think we should be considering is the dependency on hardware, like GPUs, which constrain AI advancements. While software capabilities are rapidly growing, AI's future also depends on hardware availability and efficiency, particularly as demand for computational power surges. — Leandro DalleMule, general manager, Planck / Applied Systems

AI Will Impact Payments with Personalized, Real-Time Payout Solutions

In 2025, AI will make the biggest impact in payments by driving unprecedented levels of personalization. Increasingly, consumers expect payment experiences that cater to their specific needs, whether that is real-time pay or preferred payment methods. AI can be embedded into a businesses' payout process to intelligently ensure that hundreds or thousands of recipients receive their funds in their preferred methods in real time — across digital wallets, bank transfers, and prepaid cards. — Gabriel Grisham, vice president, PayQuicker

AI Will Revolutionize Document Interaction and Content Creation

As we embark on the journey of understanding AI's impact on productivity, particularly in how we create, interact with, and experience documents, we find ourselves at an exciting crossroads. Currently, generative AI is revolutionizing content creation, enabling us to produce new material with unprecedented ease. Additionally, AI's capability to access and summarize text from images has transformed our interactions with documents, making them more intuitive than ever. Looking ahead, I anticipate a significant evolution in how we experience these documents. One of AI's groundbreaking advancements is its ability to establish a direct interface between humans and computers. While the popularity of natural language chatbots is currently capturing attention, they serve as a preliminary step in demonstrating the potential of transformer models.

In the coming year, we can expect this new human-computer interface to facilitate real-time personalization of document content. This means that interactions will become dynamic, tailored to individual preferences and past experiences, all while leveraging the most current information available. Over time, the conveniences brought by AI will become so integrated into our daily lives that they will be taken for granted, much like our constant connectivity to the internet today. — Jonathan Rhyne, CEO & co-founder, Nutrient

The Next Wave of AI Transformation

In 2025, agentic AI is set to transform business operations by driving autonomy and transparency across complex workflows, particularly in heavily regulated industries such as healthcare, finance, and energy. Acting as proactive, intelligent partners, these AI agents automate complex processes and adapt in real time, all under strategic human oversight. This integrated collaboration between AI and people enables organizations to boost efficiency without sacrificing transparency and control, ensuring compliance with robust audit trails. Placing AI at the core of competitive strategy empowers businesses to navigate and anticipate evolving challenges, with people guiding and refining AI's role for meaningful impact. — Joe Dunleavy, global SVP and head of Dava.x AI, Endava

Agentic AI Will Need Less Human Input

Agentic AI will become more dynamic and collaborative, requiring less human input. — Juan Jose (JJ) Lopez Murphy, head of data science and AI, Globant

The Rise of AI as a Digital Workforce

In the coming year, AI agents will establish themselves as a pervasive digital workforce, working in parallel to human teams. We're also likely to see the first major success of an autonomous business — a company run entirely by AI, with humans playing a minimal role, if any. This milestone will likely emerge in e-commerce, where AI-driven agents will manage customer service, inventory, logistics, and pricing, responding instantly to shifts in consumer demand and setting new efficiency standards. These advancements will drive an urgent need for AI governance frameworks, similar to HR systems for human teams, to oversee the ethical and operational management of AI agents. As AI takes a more central role, businesses will be challenged to create new standards for accountability, transparency, and performance — essentially, HR for the digital workforce. — Adriano Koshiyama, co-founder and co-CEO, Holistic AI

AI's Growing Role in Streamlining Property Management While Keeping a Human Touch

As property managers continue to face burnout and demands for efficiency from renters increase, AI is now at the point where it can be widely adopted to automate the most difficult tasks. But its successful adoption in multifamily properties will depend on the careful balance between automation with human interaction to ensure renter satisfaction. In 2025, property managers will use AI-driven solutions to help streamline administrative workflows while they can focus on what they do best — meeting renters where they are at — whether that is hosting events and tours, or helping them resolve a maintenance issue. — Virginia Love, industry principal, Entrata

AI Companies Will Be Evaluated on Ability to Deliver Tangible Results

Murmurs of an AI bubble have fed concerns about AI's long-term potential. The antidote is demonstrable return on investment. Investors and consumers alike are now demanding evidence of real-life use and tangible results. Going into 2025, AI companies will be judged on their ability to transform promises into performance and showcase the impact of their technology in the real world. This will come in the form of longstanding, satisfied customers who can speak to substantiated commercial results. These kinds of stringent expectations promise to surface the dominant players in the AI space and cut through the noise, resulting in a more navigable market for consumers. — Eleanor Lightbody, CEO, Luminance

LLMs and LAMs Will Reshape Different Sectors

As we look ahead to 2025, the integration of Large Language Models (LLMs) and Large Action Models (LAMs) is poised to revolutionize various industries. These advanced AI technologies are not just enhancing efficiency but also transforming the way businesses operate and interact with their customers. Here are some key predictions on how LLMs and LAMs will reshape different sectors:

CEOs Will Focus on Human-AI Partnerships

As enterprises expand AI implementation and automation, success in 2025 will be more dependent on a leader's ability to create a partnership between AI systems and human teams, rather than those who try to implement the newest technologies without any kind of roadmap. Moving forward, CEOs must encourage their employees to remain involved in refining AI outputs and fostering a feedback loop, ensuring that technology amplifies human judgment and capabilities (which it certainly can!). This approach, rooted in oversight, gradual integration, and continuous adaptation, will drive businesses forward. — Alon Goren, CEO and co-founder, AnswerRocket

Agentic AI Is Shifting Workflows

Agentic AI marks a transformative shift where AI can autonomously plan, execute, and verify complex workflows with genuine "agency," creating assets that enterprises can reuse and evolve. Heading into 2025, foundation model companies like OpenAI, Meta, and Anthropic are "all in" on this agentic pattern, refining AI models to independently choose actions, conduct research, apply tools, and perform iterative corrections. These advanced models promise capabilities far beyond traditional chatbots, able to drive complex enterprise processes while ensuring accuracy and alignment through embedded "verifiers" and human collaboration. We're at a peak in the hype cycle now, with the latest models able to execute agentic prompts that were previously off limits due to context windows, costs or just plain smarts.  Unfortunately, the technology has moved faster than the enterprise buying cycle so many AI-based solutions are simply calling themselves "agents" or "agentic" even though solutions have not changed. We will be seeing more true agentic workflows in 2025. — Mike Finley, CTO and co-founder, AnswerRocket

Companies Will Seek External AI Solutions

Most companies that were debating building vs. buying AI solutions will have experimented in-house with developing their own AI solutions. They'll realize they lack resources and expertise to effectively support this and begin earnestly looking at external solutions and partners. — Pete Reilly, COO and co-founder, AnswerRocket

Customer Expectations and AI

AI and predictive analytics are great for proactively identifying patterns and anticipating churn risks, so we can address issues before they become deal-breakers. AI-powered tools also are really effective in personalizing customer interactions, analyzing engagement trends, and providing well-informed recommendations across our teams, enabling customer success teams to act quickly and strategically with all the information they need across the enterprise. The advent of agentic AI will enable the Customer Success function to provide much more personalized service to customers. At the moment this approach might still appear quite "novel," but it will soon become mainstream. — Deb Ashton, SVP of Customer Experience and co-founder, Certinia

AI Shift to Outcome Models

AI will drive a transformative shift in the value model for professional services organizations through two key impacts that redefine client outcomes. First, productivity gains from Generative AI (Gen AI) will streamline knowledge-based workflows, turning tasks like document summarization, proposal drafting, data analysis, and client reporting into rapid, scalable processes — enabling services organizations to achieve in hours what once took days. Second, Agent AI will introduce a new paradigm for autonomous task management, with virtual "personas" such as Customer Success Managers and Resource Managers independently handling routine activities within well-defined guardrails for compliance and data privacy.

Together, GenAI and Agent AI will pave the way for outcome-focused business models, leaving behind traditional "hours billed" in favor of models tied to real, measurable client impacts. This model will offer significant results, yet it brings new questions: will clients view agent-driven work as a replacement for human input, or will they see it as a distinct value layer? Either way, this AI-powered transformation will reframe the provider-client dynamic, creating an adaptive feedback loop where services can pivot in real time to align with clients' evolving priorities. In this outcome-focused model, services organizations won't just fulfill project requirements—they'll become strategic partners committed to continuous, measurable value, reinforcing client loyalty by focusing on tangible outcomes and establishing themselves as essential allies in their clients' long-term success. — Raju Malhotra, CPTO, Certinia

AI and Human Expertise Redefine Client-Provider Partnerships for Continuous Transformation

As service models evolve, so too will the relationships and roles within the client-provider landscape. Professional services organizations will move decisively beyond traditional solution implementation to become strategic transformation partners — enabled by a blend of human expertise and AI support. They will lean into AI, but with a crucial twist: humans will lead strategy and complex problem-solving, while Agent AI supports them at scale by automating ongoing tasks and issuing proactive recommendations as projects progress. This AI-human collaboration will allow them to stay agile and responsive, turning one-time projects into continuous, transformative engagements. As Agent AI technology matures, services organizations will adapt to manage a flexible balance between human-driven insights and scalable AI support. In practice, this means humans will lead nuanced client engagements, discovery, planning, and strategy-setting, while AI handles standardized tasks under expert oversight. This approach will allow them to embed themselves in their clients' long-term goals, moving from providers of standalone solutions to essential partners in ongoing AI-driven business transformation. — Raju Malhotra, CPTO, Certinia

AI Evolves From Assistant to Coach

Today, AI primarily serves as an "assistant," streamlining routine tasks and boosting efficiency. But within the next few years, AI will evolve from simple assistance to active coaching, empowering professionals to acquire new skills in real time. This next generation of AI will act as a dynamic learning engine — guiding professionals in building broad skills like effective management and in-depth, role-specific skills, such as mastering Salesforce CRM or navigating company-specific onboarding. For professional services organizations, this shift will redefine how teams approach continuous learning and problem-solving within their fields. As AI evolves into a coaching role, it will become the new discovery engine—one that advances beyond static, search-based models to offer dynamic, personalized learning. Tailored to the unique needs of each organization and professional, this new AI-driven engine will create a continuous development journey. In this future, AI stands to become as foundational to professional growth as mentors and formal training programs are today, accelerating skill development and transforming career progression across the industry. As AI matures, professional services firms will not only adapt but thrive, making themselves indispensable to clients in an ever-evolving, AI-enhanced landscape. — Raju Malhotra, CPTO, Certinia

AI Will Create a New User Experience for Industrial Software

AI will increasingly serve as the front end for industrial software. Rather than navigating complicated menu systems, users will interact with AI in simple terms—asking it to perform tasks, generate insights, or design models and dashboards. As the trend is replicated in industrial settings, its benefits to business will show up as enhanced productivity, streamlined workflows and shorter time-to-value, all without heavy investment in retraining. — Jim Chappell, global head of AI and advanced analytics, AVEVA

Humanized AI Will Lead to Wider Accessibility

The trend toward natural language and voice-based interfaces will allow operators with little or no technical training to interact more closely with all types of AI. Even non-expert industrial workers will now begin to use AI to do their jobs better — without needing to understand the technology at work, whether these are neural networks or generic algorithms. Humans will become fluent in AI, taking a step closer to Industry 5.0. — Jim Chappell, global head of AI and advanced analytics, AVEVA

AI Will Do More of the Industrial 'Heavy Lifting'

GenAI will increasingly interface with both humans and other types of AI, allowing capabilities to be provided that were previously never possible.  Further, autonomous AI systems are now able to handle dynamic processes, responding to changes and disruptions near-instantaneously. With more predictable intelligent outputs, industrial operators will benefit from stable, timely production as well as improved collaboration and innovation across the value chain. — Jim Chappell, global head of AI and advanced analytics, AVEVA

Maybe AI Was the Friends We Made Along the Way

AI companionship has been a hot topic across industries, from tech and business to mainstream culture. Yet, real-world interactions with these technologies have barely scratched the surface. 2025 will change that. Picture it like an SNL skit gone viral: AI will become the centerpiece of dinner table debates, holiday gatherings, and cultural discussions. The psychological and societal impact of these new "friends" will be impossible to ignore, shifting AI from a buzzword to a core topic of how we navigate the future of human connection. — Joshua Terry, director of product management, Aura

AI in Human Clothing Will Be 2025's Trojan Horse

Legendary Chessmaster Garry Kasparov once noted that a human working with a computer can consistently outplay even the best chess computer. As AI vs. AI evolves in digital security, this same pattern will be key. AI that mimics human behavior will need systems that combine human expertise with advanced algorithms and data. Take financial institutions that use AI to flag unusual transactions. As attackers increasingly mimic human spending patterns, the AI will rely on human input to spot trends, train models, and improve detection. As the line between genuine and fraudulent blurs, we'll see a major shift in security — one that's already starting. — Joshua Terry, director of product management, Aura

The Rise of 'Bring Your Own' AI Models

In 2025, the trend of "Bring Your Own" (BYO) AI models is poised to accelerate, letting businesses integrate their own data assets and custom-trained AI models into third-party platforms. Rather than relying on pre-configured solutions, companies will be able to leverage proprietary large language models (LLMs) or data lakes across various tools, particularly in complex martech stacks. This shift is expected to streamline workflows by reducing the redundancy of re-training models across multiple applications, enabling a seamless, customized AI integration that adapts to unique business data and insights. This will be particularly appealing to companies that have invested heavily in tailored AI models based on their customer and product data, fostering more personalized, efficient, and scalable use of AI across enterprise applications. — Karl Bagci, head of information security, Exclaimer

Regulators Will Provide More Guidance on AI Use in Financial Services

The growth of innovative AI solutions merging customer data and financial service products and services through an unleashing of Open Banking technology product breakthroughs will break down barriers in financial services and provide helpful solutions for millions across the globe, but a couple of personal data misuses in the papers will spur regulators in the UK, EU, US, and Singapore to take action and provide more specific guidance on AI use in the sector. — Vaikkunth (Vaik) Mugunthan, CEO/co-founder, Dynamo AI

Expect a Surge in Development of AI Agents

By 2025, AI Agents — intelligent systems capable of making decisions and executing tasks autonomously — will revolutionize industries, offering far greater value than today's chatbots. However, their complexity will pose challenges in evaluation, debugging, monitoring, and optimization. This will create a significant market opportunity for companies focused on making AI Agent deployment safer and more manageable. As a result, we can anticipate a surge in both the development of AI Agents and the emergence of supporting technologies and services that help organizations harness their power while minimizing risks, giving those who succeed in this space a competitive edge. — Vaikkunth (Vaik) Mugunthan, CEO/co-founder, Dynamo AI

AI Agents Will Transform Automation

The world of AI is evolving to the next generation of automation, AI Agents. As with everything AI can be used for good or bad purposes. AI Agents will evolve in 2025, to become a real-world production ready capability where these intelligent systems will make decisions. There will be situations where the AI agent will run with some human oversight and in other cases without. Their ability to react to unforeseen scenarios and make recommendations/decisions 24x7 365 will be a game changer. This is especially true in the world of IT security where the increasing rate, speed, agility and volume of attacks is the next challenge for security leaders and their SOC teams. Visionary security executives are already expressing the need to be able to defend at machine speed. Since the attackers will also be leveraging autonomous AI agents, this will lead the way for them to attack intelligently and these malicious actors to also try to exploit AI/ML front ends as new attack vectors as well. For developers, these agents will be needed for constant updates to the ever-growing needs of the distributed software, potentially sacrificing security for speed and efficiency but at the same time increasing business risk. — Paul Davis, field CISO, JFrog

The Shift from GenAI to Fine-Tuned Multi-Model Integrations

After the excitement of multi-model AI integrations in 2023 and 2024, which highlighted their impact and cost-benefit, the next phase will focus on fine-tuning these integrations. In 2025, enterprises will move beyond generic use cases — such as generating code, images, and text — and begin fine-tuning AI models to meet their unique needs. Rather than relying on massive, all-encompassing models, organizations will leverage machine learning to identify top performers and use that refined knowledge to tailor AI, ensuring more secure, efficient, and specialized solutions. — Danny Allan, CTO, Snyk

The Rise of Security-Focused AI Models in 2025

As enterprises increasingly adopt coding assistants and autonomous systems, security must move from an afterthought to a priority. In 2025, AI models trained on generic, high-volume data often suggest common but insecure solutions, leading to complex systems and vulnerabilities. To address this, businesses will shift to multi-model integrations that prioritize security by focusing on top performers with a track record of producing secure, efficient code. This will lead to the widespread adoption of fine-tuned AI models that not only drive productivity but also deliver robust, secure systems. — Danny Allan, CTO, Snyk

AI Skepticism Will Give Way to AI Confidence

Despite continued executive urgency to incorporate AI tools into business operations, more than two thirds of desk workers still say they've never used AI at work, and concerns about accuracy are still barriers to reliance (just 7% of desk workers say they consider the outputs of AI completely trustworthy for work-related tasks!). This points to an urgent need for businesses to more closely connect the power of AI to users' daily work, in the unique ways that they require, and in the most dead-simple way possible. In 2025, we'll see the barrier wane as users work side by side with agents for common tasks like automating project tasks, new hire onboarding, generating content, or managing IT incidents. Agents' advanced reasoning and abilities to make decisions and take action will transform how every user works and how they engage with each other and customers. 2025 will be the year desk workers grow more confident with AI and businesses will see even greater adoption and ROI on their investments. — Rob Seaman, Slack chief product officer, Salesforce

As AI Matures, Massive LLMs Will Be Replaced by Ones That Are Purpose-Built 

The notion that bigger is better will fade as organizations begin to recognize the pitfalls of excessiveness with AI models. Narrowing the sources of data and scope of information allows the LLM to become highly specialized around what best serves an individual organization. Refining the focus ensures the output is relevant and doesn't waste sources on unnecessary knowledge or capabilities. For example, an apparel retailer doesn't need its LLM to know about agriculture or medical research — training the AI to understand nuances about textile supply chain mechanics and the specific payroll processes of the organization is more beneficial. And privately hosting the dedication LLM allows for increased security, less bias, and enhanced accuracy. — David Lloyd, chief AI officer, Dayforce

AI's Role in Evolving Telehealth

Telehealth is now a permanent fixture, with AI expanding its reach and capabilities. AI's role will focus on securing AI-powered video consultations, ensuring data privacy, and enabling seamless integration with patient data systems. This support will enable healthcare providers to scale telehealth services securely and efficiently, meeting the growing demand for accessible, remote healthcare solutions. — Shash Anand, SVP of product strategy, SOTI

Seeing Through the AI Buzzword Looking Glass

2025 will usher in a pivotal shift for AI, moving beyond the rapid growth and hype of recent years toward a more grounded and practical phase. Companies will increasingly realize that building AI solutions in-house is far more complex than anticipated, prompting a shift toward buying and integrating established technologies. Buzzwords like "agentic AI" will continue to gain traction, but the broader narrative will evolve. Instead of focusing on catchy terms, discussions will center on addressing real-world challenges—closing learning gaps, overcoming deployment hurdles, and delivering measurable outcomes. This shift signals a maturing industry focused on sustainable impact and long-term trust in AI systems. — Assaf Melochna, co-founder, Aquant

Generative AI Is the New Customer Experience Differentiator

Generative AI is moving out of the shadows and taking center stage as a competitive differentiator. Customers today want to know that the brands they trust are leveraging cutting-edge technology. By showcasing generative AI as an asset and a core part of CX strategy, we can enhance transparency and build deeper trust, showing customers the real value behind every interaction. — Assaf Melochna, co-founder, Aquant

The AI Hangover Is Settling In — Living (and Working) with the Reality of AI

AI may have grabbed the headlines in 2024, but in 2025 organizations are going to get real about how they want to use AI — and the realities of implementing it. AI today can perform some impressive feats—generate artistic images, answer open-ended questions — actions formerly the province of humans alone. But it can also do a lot of the more "boring," tedious manual tasks that bog our day-to-day work down. In 2025, organizations are going to end their AI exploration phase to instead take a deep, realistic look at their need for the technology and how it will meaningfully help their business and customers. And they'll find that their best minds will not be replaced by AI, but will see how well AI can amplify their expertise. While AI doesn't create the idea, AI can help make the idea a reality faster. We're going to start seeing businesses tapping virtual agents and copilots for the tedious work while letting humans do what they do best — be creative. — Skip Levens, product leader & AI strategist, Media & Entertainment, Quantum

Growing Up in the Age of AI — What's Real, What's Not?

What happens when almost every piece of "born digital" media seen on the web and social media meets an avalanche of readily available generative AI tools? It means almost everything you see in your digital day could have been generated by AI — and inherently untrustworthy. The effects of this today might provoke a laugh or a gasp for a relatively crude implementation (why do AI images always have the wrong number of fingers?) — but the implications of pervasive and increasingly higher quality gen AI tools will be far reaching. Every business, every walk of life, every institution will need to evaluate their communication strategy, transparency in using these tools, sources of their training data, and more as the technology matures. — Skip Levens, product leader & AI strategist, Media & Entertainment, Quantum

Considerations Around AI's Growth & Adoption in 2025

There are concerns that the hype around AI will fade. Once organizations realize that the actual impact that AI can create is less than what they expected, they may scale their programs down. Therefore, it's important to be realistic when it comes to AI's expected impact on an organization and to demonstrate quick wins to prove value. Adoption of AI is always one of the key issues in any AI transformation. Creating a culture that embraces and understands AI will be necessary to overcome any resistance. — Christoph Wollersheim, consultant, Egon Zehnder

GenAI Will Evolve and Agents Will Mature in 2025 as Dark Data Is Uncovered

Once companies understand GenAI is not a Swiss Army knife that can solve every business problem, companies will find applications for which GenAI is well suited, aligning with business goals, particularly as agentic AI is more widely adopted. As companies focus on preparing data sets for GenAI, this will uncover "dark data" across the organization — in sales, customer support, marketing, finance -- which will increase the accuracy of LLMs and lead to an increase in ROI. — Scott Francis, technology evangelist, PFU America

A Numbers Game: GenAI Moves Beyond Text

Enterprises continue to leverage GenAI in new ways to guide their businesses. But they're finding out firsthand that the dominant GenAI technology (LLMs) wasn't built for that. LLMs were generally designed for text-based AI applications like content generation, chatbots, and knowledge bases. They are not effective in scenarios that require deep numerical predictive and statistical modeling to predict how a given variable will change over time based on one or more input variables (aka regression tasks). Gartner recently broke it down: "Use cases in the categories of prediction and forecasting, planning and optimization, decision intelligence, and autonomous systems are not currently a good fit for the use of GenAI models [LLMs] in isolation." At a high-level, this means that LLMs aren't great at fundamental business planning use cases, which cover things like logistics, marketing, staffing, investing, product development, and all sorts of other areas. Those applications require modelling of enterprise-specific, tabular and time series data that span key areas of the business, including people, products, sales, and budgets. The industry will respond to this gap in 2025. This year, more GenAI technologies will emerge that are engineered specifically for modelling structured numerical and statistical data rather than just text. These technologies will allow enterprises to use their tabular business data to make better decisions, minimize risk, and boost efficiencies. — Devavrat Shah, CEO, Ikigai Labs

AI Education Becomes C-Level Priority

In the era of AI, traditional white-collar workers worry about job security, while employees with strong tech and AI expertise are in major demand. It's assumed that the former will be replaced by the latter. But reskilling workers is easier than replacing them: That's why this year we'll instead see enterprises invest aggressively in AI education for their existing employees. LLMs have already shown how everyone can effectively wield AI. Rather than compete over a small pool of AI experts, major companies will re-train workers to leverage the technology. AI upskilling will soon become a C-level priority. — Devavrat Shah, CEO, Ikigai Labs

AI's Impact on Healthcare in 2025

In 2025, the industry and world will slip into Gartner's "Trough of Disillusionment " regarding AI in healthcare. Currently, every organization is racing to claim to have AI under the hood, but the fact remains that only a few really do. AI research and implementation is very expensive and time-consuming, limiting only the largest medical centers with vast resources to being able to investigate true, AI-powered solutions that can help with better outcomes. Significant improvement in the quality of the data training the AI algorithms is also needed. There currently exists siloed data, gaps in data, and simply bad data everywhere. Until it can be improved in a meaningful way, AI will not have a significant impact on healthcare in 2025. — Eric Demers, CEO, Madaket Health

Special-Purpose LLMs Will Give GenAI a More Strategic Role in Modern IT While Empowering Human Ingenuity

Generative AI will move from novel data synthesis to domain subject matter expertise in AIOps, driven by the combination of expertise encoding, retrieval-augmented generation, and LLMs. Special-purpose LLM deployments will effectively model expert decision-making across multiple ITOM and service assurance process areas. This strategic shift will mark a turning point within today's IT organizations, with generative AI playing a more streamlined and specialized role in operational areas, freeing up more time for innovation. — Casey Kindiger, CEO, Grokstream

Gap Between AI Leaders and Laggards Will Be Massive

Top companies are moving beyond just testing out AI use cases — they're making AI a central part of their business strategy. They're starting to see that AI's real power isn't in one-off projects but in how it can completely transform the way they operate. At the end of the day, succeeding with AI is just as much about having a bold, clear vision as it is about the technology itself. By 2025, the gap between companies leading in AI adoption and those lagging behind will be monumental. While not every AI promise will be fulfilled, the speed of innovation, levels of investment, and widespread business buy-in are unprecedented.  Companies that take bold steps with a strategic AI vision today will define the next era of leadership, leaving those that hesitate to play catch-up in an increasingly competitive landscape. — Dan Priest, US Chief AI Officer, PwC 

AI Will Revolutionize Physical World with Robotics Across Key Industries

While 2024 was a year for substantial advancement of AI for the digital world, I expect that 2025 will be the beginning of the proliferation of AI for use in the physical world. We will see the continued maturation and implementation of AI technology for robotics across complex, unstructured, and dynamic environments. Industries such as manufacturing, warehousing and logistics, aviation, energy, and aerospace will gain equitable and affordable access to AI solutions for robots that are easy to deploy and manage regardless of organizational scale. This will drive an acceleration in automation, specifically for tasks that historically have been too complex to automate, enabling more efficient and accurate execution of complex tasks by robotic systems, while keeping frontline workers safer. — Ben Wolff, CEO, Palladyne AI

AI's Role in Legal Tech Will Expand

AI will remain a dominating influence in the legal tech space in 2025. We can expect the rise of intelligent agents as AI's capabilities advance beyond singular tasks to orchestrating more complex workflows. At the same time, there will be an increased emphasis on improving data accuracy and quality to boost the value created by AI. This will help forge better systems-of-record, followed by a decreased focus on point AI tools. Advancements in knowledge management will also succeed static content with dynamically generated, on-demand insights, again orchestrated by intelligent agents. — Tom Dunlop, CEO, Summize

Big Tech Bets on GenAI — Will the Risk Be Worth the Reward?

Recent earnings reports from major players like Meta, Google, Amazon, and Microsoft revealed a spike in quarterly capital expenses—capital being invested in land, data centers, networking, and GPU. The payback from the capital is not clear, but the reports indicate that the payback time could take up to 15 years. This is a staggering amount of capital and an extraordinarily risky bet. What's more, this investment is not coming from the venture capital community; it's a Balance Sheet item for these companies, and the cash is coming from their reserves. Why is Big Tech making such risky investments? Simple: because they cannot afford not to. If they don't make the investment, they will be shut out of the race. We are witnessing a market transition: If you look at the last 30 to 40 years in the tech industry, we have never seen capital investments at this scale. GenAI is going to become the next platform and to play in that, companies must make these kinds of capital investments or risk becoming irrelevant. — Ratan Tipirneni, president and CEO, Tigera

Manufacturers Will Lay the Groundwork for True GenAI Implementation

Once the potential of GenAI to transform manufacturing became clear, it became impossible to look back. GenAI will optimize production processes, accelerate product development, enhance supply chain performance, and improve decision-making for manufacturers. The next two years will be pivotal as manufacturers focus on laying the groundwork for GenAI's broader implementation, which we expect to scale over the next three to five years. This includes developing well-informed, data-driven strategies to manage the influx of information GenAI will generate and ensuring teams are trained to harness its capabilities effectively. By building strong frameworks and aligning existing systems, manufacturers will be prepared for the transformational changes GenAI will bring. — Eddy Azad, CEO, Parsec Automation

Data-Centric AI Takes the Lead

Better data, not bigger models, is the real path forward. The coming year will see an industry-wide embrace of data-centric AI, where improving dataset quality directly improves model results. Companies will also train their own small language models, as opposed to fine-tuning large models. This means companies will invest heavily in high-quality, domain-specific datasets, automated data cleaning, and monitoring, leading to smarter models that rely less on sheer scale and more on nuanced understanding. — Luca Antiga, CTO, Lightning AI

Rethinking Collaboration with AI-First Platforms

Companies are waking up to the challenges of scaling AI across global teams. The next generation of platforms will treat AI as a first-class citizen, allowing seamless data integration, model development, and deployment workflows. These platforms will make collaboration between technical and non-technical teams second-nature, accelerating deployment and democratizing AI within organizations. — William Falcon, founder and CEO, Lightning AI