Tractica Report: Natural Language Processing for the Enterprise

Artificial intelligence (AI) technologies such as machine learning (ML) and deep learning (DL) are dazzling in and of themselves, but believe it or not, leveraged in isolation, they are limited in their potential. These technologies do not interpret data by themselves: they are tied either to deterministic, hard coded software programs created by humans or they are linked to a form of artificial intelligence that can interpret human language into a form ML and DL algorithms can understand. The umbrella term for this gateway AI technology is natural language processing (NLP). Other terms associated with NLP include natural language understanding, natural language generation, voice recognition, and speech recognition.  Tractica considers all of these terms and technologies as part of the family of NLP technologies, and each of them are addressed in some fashion in this report.

The ultimate promise of NLP technology is intent-based computing. When machines can understand and communicate with humans in natural (human) language, it democratizes data science, enabling humans to use common everyday language to complete a broad range of tasks from the simple and mundane such as auto-completing an online form to the most complex, such as writing software or optimizing a data network. NLP is a critical technology for extracting insights and analysis from a vast amount of previously unindexed and unstructured data; mining video and audio files, emails, scanned documents, and more.

Market Drivers

There are several market drivers for the adoption of NLP-fueled use cases, both from the supply and demand sides.

The Interface For Connected Life

With the growth of connected, intelligent devices, our lives are becoming increasingly computing-centric. We will interact with connected devices nearly constantly, from smartphones, smart speakers, and wearables to eventually autonomous cars and smart cities.  Menu-less web navigation will become increasingly important. For all of this computing to take place, computers are required to understand human inputs in the form of human (natural) language, complete tasks, and communicate back to humans in NL.

Tapping Unstructured Data, Powering Big Data

The promise of Big Data, for things like predictive analytics, advanced security, competitive intelligence, business efficiencies, and more have remained largely unfulfilled. The world is producing exponential amounts of data, but most of it has been rendered unusable. By most estimates, 80% to 90% of the data in existence is unstructured. This data is locked in scanned documents, email, video, audio, and social media. NLP has the potential to unlock the power of Big Data because it can interpret both structured and unstructured data.

Reducing Costs, Increasing Efficiencies

NLP can influence cost savings and new revenue generation. Although each use case varies, cost reductions associated with NLP tend to fall into the following categories:

Opportunities for new revenue generation associated with NLP tend to fall into the following categories: 

Market Barriers

Understanding Context

The biggest market barrier for NLP use cases is the low level of accuracy for current NLP technology. Machines have a difficult time understanding humans. The reason for this is context. It is difficult to understand the difference between what someone says and what they mean, and computers are very literal. Understanding language requires the highest form of intelligence. It requires the understanding of location, tone, implied reference, and history. Sarcasm and emotion are hard for computers to interpret. Conversational and relational history is also hard for computers to interpret.

Lack of Clean, Accurate Data

Most NLP systems require statistically valid, clean, and accurate data. As with any information system, bad data will result in bad assumptions and predictions, but because many NLP systems program themselves, clean data will be even more critical. 

Use Cases

NLP is increasingly being used in a wide range of applications across many industries. Hybrid use of NLP with ML and DL will grow dramatically. Section 3 contains summaries of the 43 key hybrid NLP-ML/DL use cases, ranked by overall revenue over the forecast period of 2017 to 2025, in descending order.

Key Industry Players

This report mentions more than 60 companies, including 43 that are profiled in Sections 3 and 5. The sheer number of companies involved in the NLP market ecosystem is a good indicator of the interest in NLP and breadth of applicability of the technology across a number of industries and sectors.

Market Forecast Highlights

Tractica forecasts that annual revenue for NLP software will increase from $813 million worldwide in 2017 to $9.1 billion in 2025, representing a compound annual growth rate (CAGR) of 35.3%. Total annual revenue for NLP, taking into account software, services, and hardware, will increase from $3.2 billion in 2017 to $43.3 billion in 2025, at a CAGR of 38.6%

As with many software technologies, the most robust adoption in the early years will occur in North America and deployments will subsequently spread to the rest of the world, although, for some use cases, less developed regions may leapfrog more mature markets.

This report forecasts hardware and services revenue driven by NLP. The hardware forecasts are segmented into separate forecasts of the demand for central processing units (CPUs); field-programmable gate arrays (FPGAs); application-specific integrated circuits (ASICs); graphics processing units (GPUs); networking products; and data storage devices.

NLP will also generate significant demand for professional services, such as installation, training, customization, application integration, and maintenance.

Conclusions and Recommendations

Conclusions for this report have changed very little since Tractica’s previous report on the subject in 2017. There is consistency in what is coming from NLP, though as with all artificial intelligence (AI) technology, Tractica’s forecasts for adoption show accelerated growth over forecasts of just a year ago.

The democratization of computer interactions enabled by NLP will allow non-technical workers, in essence, to become computer programmers. NLP has the potential to allow non-technical workers to describe a problem in NL and instruct computers to solve the problem.

NLP unlocks the promise of Big Data due to its ability to read and interpret unstructured data. In this sense, NLP is the key to much more sophisticated and usable analytics.

NLP technology will appeal to a broad range of industries, workflows, and mechanisms, particularly those that require humans to interpret large sets of structured and unstructured data. NLP has the potential to reduce human costs and increase operational efficiencies.

It is unlikely that NLP will reach a point in the next 10 years or perhaps ever, when it will be able to match humans for understanding language. Context is too nuanced of a concept for computers to grasp. Because of this, there are practical limitations to the use cases for NLP.

The most successful NLP use cases over the next 3 to 5 years will be those that do not require context to succeed. Practical applications, such as contract analysis, compliance monitoring, business and market intelligence, and streamlining healthcare, will lead the market for NLP services.