CONVERSPEECH LLC has been developing software for R&D since 1997. Located in Palo Alto, California, in the heart of Silicon Valley and headed by Colleen E Crangle, it has expertise in specialist computer technologies applied to problems that affect our world. Launched and supported by Small Business Innovation Research awards from the National Institutes of Health (NIH), it has been at the forefront of NLP and machine learning research applied to healthcare.
CONVERSPEECH LLC is a limited liability company registered in the state of California. For further enquiries about the company please use our Contact Form.

Our Team

David Gleeson, Associate, audio engineering

Michael McTear, Associate, spoken language technologies


Our signature WhiskeyTM software is provided along with a service that delivers a questions-based assessment of your website.

Your problem: What do people want to know about your service or product and can they get those questions answered at your site? Our solution:

  • 2-hour consultation with your marketing team.
  • Compilation of 30-100 questions covering a proprietary blend of cognitive domains.
  • Delivery of a WhiskeyTM text analysis of your site, which includes:
  • Statistical analysis of your site's alignment with your client/customers' expressed needs.
  • Stratification of these results by demographics such as geographic location or gender or custom characteristics.
  • Analysis in terms of how information rich your site is and its perceived difficulty.

Basic package described here costs $1,560

How to order: please send an email to whiskey at converspeech.com with your best contact information.

Sometimes the questions are complicated and the answers are simple.
Dr. Seuss.


Natural Language Technologies

For computers to serve humanity adequately they must be able to work with human language. Only then will we have ...

  • computers that can analyze large collections of written text. From scientific articles and electronic health records to emails, twitter and chat streams, we encode our experience of the world in language.
  • a world of Connected Health, where bio-sensing technology and data analysis are complemented by accurate and sensitive human-machine communication
  • computers that do natural language processing (NLP) as seamlessly as people do, conversing with understanding

Data Science Technologies

Data derive from a variety of sources. Imperfect, mixed, varied, and inconsistent, data's first challenge is to accept its inherent messiness. The second challenge is to accept its uncertainties.

  • Automating the import of data from a variety of sources is often a necessity: OCR-ed data, annotated records, web data, sensor-based data, and crowdsourced text and categorical data. Such varied and error-prone sources frequently require cleaning, normalization and aggregation of the data.
  • Attitudinal data analysis gives insight into human emotions and attitudes. These essentially probabilistic human phenomena require predictive analysis at the level of the individual (how likely is someone to react in a certain way to specific events) and at the level of the group (how do people with this characteristic tend to behave).
  • Machine learning allows the computer to make sense of data with or without human input.
  • Supervised learning, unsupervised learning, and reinforcement learning are three major approaches. Probabilistic, rule-based, and knowledge-rich algorithms must be designed as deployed as demanded by the data and our 30-year history in machine-learning development.

Neurocomputational Technologies

The mind is this century's scientific frontier. Exploring it with intelligent use of technology will give us ...

  • brain-computer interfaces as communication aids
  • new tools for understanding language, emotion and the mind, yielding new information about human behavior

Why these technologies matters to society

These technologies have the potential to transform the way we solve social problems.

    Data derived from government and commercial sources, including social media sites, mobile phones and transaction data, are offering new ways to understand and predict human behavior. Can this potential be exploited for the good of society? What opportunities are presented by big data? What are the dangers?

  • Nowhere is the data problem more technically challenging than with text. Words, utterances, documents contain information vital to commerce, science, and public governance. The insights afforded by the analysis of large volumes of textual data are yet to be fully understood and appreciated.

  • Computational neuroscience is poised to give us the most significant window ever into the human mind and human behavior.


A round-up of projects in health data science, plus data resources from the National Institutes of Health


Our signature WhiskeyTM software-with-a-service gives a questions-based assessment of your website.

See further details here.

Business case development and strategic technology assessment

Opportunities associated with natural-language processing (NLP), machine learning, emotion detection, cognitive neuroscience technology