With digital transformation driving a new imperative for processing unstructured data across multiple industries, artificial intelligence solutions incorporating natural language processing (NLP) are now considered business critical.
Language is increasingly being recognized as THE delivery mechanism in the artificial intelligence arena, and language applications—as we’ve seen with the buzz around large language models (LLMs), like ChatGPT, are a key use case because those applications need more than just a superficial understanding of text.
Therefore, it’s no surprise that spending on AI technologies like NLP is expected to increase this year. This was certainly one of the key takeaways from our recent survey of NLP practitioners in the US and Europe, where more than three-quarters of organizations surveyed expect to spend more on NLP projects in the next 12-18 months. It’s also consistent with what we’re hearing in the marketplace.
Fortune Business Insights predicts that the NLP market will be $127 billion by 2028. Forrester sees NLP as a technology that is at the forefront of delivering on the promise of AI and ROI not “one day,” but today. According to Brian Hopkins, VP Emerging Tech Portfolio, Forrester Research, “NLP benefits are almost immediate. According to Forrester’s 2021 data, 70% of data and analytics decision-makers whose firm is adopting AI say they expect their firm to use natural language technologies…”
Let’s take a closer look at the major findings from our research and the trends for 2023.
Major Trends in NLP
The main findings of our research offer a glimpse into the state of Natural Language Processing in 2023:
- Companies want to be able to address multiple use cases: 71% use a natural language platform that supports multiple use cases
- Accuracy is a priority, and a challenge: 67% of businesses with NLP models in production for 5+ years still deal with accuracy challenges
- Efficiency is the benchmark for ROI: 61% of teams measure ROI based on efficiency improvements
- Explainability of models is important: 74% say they consider how responsible (explainable, energy efficient and unbiased) each AI approach is when selecting their solution
- NLP workloads are growing: Four in five organizations surveyed have NLP models in production, with tens of thousands of documents being processed by each organization monthly.
- Adoption challenges persist: Building the business case for NLP projects, especially in terms of ROI, is a challenge facing 37% of businesses surveyed.
Key Takeaways for Businesses
It’s clear there is a substantial appetite for operational efficiencies to reduce costs, drive growth and gain a competitive advantage delivered by using NLP, with only 1% of our respondents not actively considering NLP business use cases, and 77% expecting to increase NLP spending. So, what do these trends tell us?
Hybrid approaches work best
The vast majority of companies surveyed for this report are taking a multi-solution approach to NLP.
Machine learning and symbolic are often pitted against each other as mutually exclusive options to natural language understanding. This has forced organizations to compromise in one way or another. Recent technology advances have resulted in the availability of a hybrid approach, where organizations can use multiple techniques including machine learning (ML), deep learning (DL), symbolic approaches and even leverage large language models (LLMs) in tandem, enabling them to realize the core benefits of multiple approaches. Teams often need the flexibility of a hybrid approach that integrates multiple techniques to achieve the success metrics most valuable to each use case, such as scalability and accuracy, and explainability.
In fact, a hybrid approach that combines multiple AI methodologies like symbolic AI and ML is increasingly recognized as a “best of both worlds” choice for natural language AI applications. A recent Forrester report praised a hybrid approach: “For a future fit, adaptable NLP solution that is out of the box, only requires moderate support, and can continuously learn and self-improve, look for an NLP solution based on hybrid AI.”
Businesses know what they want to achieve
Four in five (81%) of respondents said their company has a clear vision of how to use AI natural language projects, with 34% strongly agreeing with the statement.
Among the business use cases currently using or considering NLP, data protection and governance (think GDPR, PII) and knowledge management and classification were the top choices. While the technical use cases for NLP were being deployed for extraction / named entity recognition, intent classification / chatbots, and natural language generation.
Despite the potentially endless range of potential use cases, organizations already have clarity on what they want NLP to deliver. Companies should hedge against upcoming AI regulatory requirements for transparency and explainability and ensure that they adopt responsible AI approaches sooner rather than later. This will be especially critical for integrations with generative AI applications.
Challenges will be overcome
Our data shows that only 1% of current NLP practitioners report encountering no challenges in its adoption, with many having to tackle unexpected hurdles along the way. However, as NLP becomes more ubiquitous, the top NLP adoption challenges—costs, building the business case, and lack of in-house expertise—are all within reach given the right guidance. We believe staffing skill levels will increase and existing models will be shown to increase efficiency, reduce risk, deliver competitive advantage and cut costs. In addition, the challenges of aligning stakeholders, demonstrating ROI and perceived cost barriers will recede, and organizations will move to rapidly implement NLP solutions.
There’s ambition for the future
Across the board, no matter what stage they are at on their NLP journey, businesses are overwhelmingly looking to increase investment – and grow the number of NLP models in production over the next 12 to 18 months. With such rapid expansion will come the need for deeper understanding of which technology can deliver the most accurate, repeatable and responsible approaches.
As always, we believe that the best path starts with the problem you want to solve. For most, if not all, business use cases accuracy is critical and there’s no taking chances with your data. That’s why the expert.ai platform offers a single environment that combines multiple approaches (ML, DL, LLMs, symbolic, knowledge graphs, and now, GPT) and, the domain-specific and proprietary data that characterizes most businesses, with humans in the loop.
Want to know more about the trends driving NLP investment and innovation? Join us Thursday, February 16 for a deep dive and discussion into the findings! Reserve your spot today!
