Emotions: The Next (but not new) Frontier in Artificial Intelligence & Cognitive Computing

May 20, 2017

How you “feel” has never been more important in tech. This year, a number of companies are showing an interest in emotions, from Apple’s purchase of Emotient for interpreting emotions from facial expressions in January, to IBM ’s release of new Watson APIs for analyzing emotion and tone in text.

We get it. At Expert System, we’ve always believed in the value of a deeper understanding of text. While meaning and context can tell us so much, we’re leaving a lot of valuable information on the table if we’re not looking at one of the key things that makes us human. So, it’s natural that emotion is the next frontier for conquering, albeit one whose subtleties and nuance make it even more difficult to determine with great accuracy. It’s a challenge we’ve been working on for many years, leading up to the release of a couple of our most innovative products:

  • Since 2011, with our spinoff ADmantX, we provide an API that detects and extracts 95 emotions in content, including those for intent and behavior. Applied in digital advertising, these are especially effective for anticipating user behaviors and serving up the most emotionally appropriate (and contextually relevant) ads in real time.


  • Among the rich set of data provided by Cogito Intelligence API, released two years ago, our Writeprint feature goes beyond the evaluation of tone to analyze a document based on the level of difficulty to read it, the writer’s vocabulary level and identify/analyze slang and other grammatical features of the text.

Understanding emotion in content has a range of uses, from the usual business intelligence and customer service applications, to aiding in medical diagnosis; and those are just some of the earliest applications of the technology.


So, we welcome IBM and others to the world of emotional analysis!