Software Development Productivity Crisis Costing Companies Millions

A new study from software delivery platform provider Harness reveals a significant productivity crisis in software development.

The Harness State of Software Delivery 2025 report surveyed 500 engineering leaders and developers, highlighting critical challenges in modern software development practices.

While artificial intelligence (AI) can bring benefits, the research reveals significant challenges in security and governance as organizations integrate AI into their software development lifecycle (SDLC). Even more worrisome is the fact that organizations are facing substantial financial losses due to inefficient processes and manual tasks.

Key Findings

The Harness State of Software Delivery 2025 report found that:

• Organizations lose $8 million in productivity per 250 developers annually.

• 78% of developers spend more than 30% of their time on repetitive tasks.

• 92% of developers report increased risk of problematic code reaching production.

• Only 48% use approved development tools.

• 60% lack formal processes for code assessment and security validation.

"What surprised me the most is how history is repeating itself," Nick Durkin, field CTO at Harness, told ITPro Today. "Despite all the lessons we've learned from the struggles with shadow IT and the early days of cloud adoption, we're making the same mistakes with AI."

Related:6 Software Development Trends to Watch in 2025

Durkin noted that more than half of DevOps teams are using AI tools that aren't approved by IT, which shows we haven't built easy on-ramps for engineers to adopt these tools securely and effectively.

Developers Spend Too Much Time on Repetitive Tasks, and AI Isn't Helping

Another somewhat surprising finding for Durkin was the amount of toil that developers still exert on manual tasks.

"It's staggering to think that developers are still spending so much of their time on manual, repetitive tasks — work that doesn't drive innovation or add value," Durkin said. "It's clear many companies are investing in the wrong areas instead of solving these long-standing productivity problems."

As to how organizations can improve the situation and reduce the amount of manual, repetitive tasks, Durkin has a few ideas. In his experience, the main issue is misplaced priorities and investments. 

"Too often, organizations focus on flashy AI tools that don't address core problems, like reducing toil," he said. "These tools might look impressive, but they don't solve the real workflow bottlenecks — and sometimes, they even take away the parts of the job developers actually enjoy."

Related:Top 10 DevOps Stories of 2024: Key Insights and Innovations

Durkin added that recently many AI developer tools such as code assistants aim to speed up development time. Yet despite that, many developers still spend a significant portion of their day on tasks they don't enjoy, like babysitting deployments or waiting for tests to finish.

"The key is leveraging AI to target the real inefficiencies in an organization, not just the most visible ones," Durkin said. "Teams need to take a step back and identify where they're actually losing time and energy. Without that clarity, it's impossible to choose or deploy the right tools effectively."

The Cost of Developer Burnout and What Enterprises Can Do 

The report highlights a critical productivity crisis, with organizations losing approximately $8 million annually per 250 developers.

"Burnout is expensive — plain and simple," Durkin said. "When developers feel unsupported or overwhelmed by unnecessary toil, they eventually leave."

He added that replacing a burned-out engineer isn't just costly — it's disruptive. Losing quality engineers puts additional stress on the rest of the team and often exposes deeper structural issues within the organization.

To fix that situation, Durkin recommends that companies need to make engineers' lives easier by empowering them to work on what they love, not the tasks they procrastinate. Burnout often stems from organizations prioritizing flashy but unimpactful initiatives rather than simplifying workflows and reducing toil. This often happens when companies invest in the wrong tools and processes.

"By focusing on what truly matters — eliminating inefficiencies and empowering engineers to do meaningful work — organizations can retain top talent, boost morale, and avoid the significant costs associated with burnout and turnover," Durkin said.

Security and Quality Concerns with AI

 Despite AI's promise, the technology introduces new complications. The survey found that 67% of developers spend more time debugging AI-generated code, while 68% report increased time resolving AI-related security vulnerabilities. More concerning, 59% of developers experience deployment errors at least half the time when using AI tools.

Adding to the risk, 60% of organizations lack formal processes for assessing AI-generated code vulnerabilities or errors, and 58% don't provide clear guidance on low-risk AI use cases.

"Without formal evaluation processes, many organizations don't even know where to start when it comes to measuring success," Durkin said. "They chase quick wins instead of focusing on meaningful, long-term outcomes — it's like opting for a fad diet instead of building sustainable healthy habits."