Building on the insights from earlier surveys in 2020-2021, we’ve recently conducted a more updated industry survey of firms that use natural language systems in production.

This report leverages a “contrast study” approach, by including organizations which have a history of leveraging NLP systems as well as those which are just beginning to plan their approach, then comparing the different between their priorities and practices. Within the past year this field has seen rapid evolution in the use of deep learning models (e.g., “transformers”) for commercial applications.

Subsequent to the pandemic and other events in 2020, there have been increased needs for detecting fraudulent news reports, with related NLP methods for key point analysis and argument support moving into mainstream commercial applications.

The pandemic also highlighted more urgent needs for applying abstractive summarization to make sense of the torrent of scientific publications, for example with CV-19 research summaries making the least medical news more “human scale” for overworked healthcare workers.

Our survey results explore this evolving landscape for NLP to help spotlight the latest trends.