What if you could conversationally locate complex information during your work tasks? Say you’re an engineer looking up design parameters from within tables and text. Or an insurance sales person wanting to locate a property insurance plan that conforms to several criteria.

You could be a plant manager searching across a year’s worth of product performance data. Or a Vice President of Sales reviewing revenue streams from a specific client. Isn’t there always information you need quickly – and looking through tables, reports and writing SQL queries is not your idea of fun?

Don’t we sometimes wish we were living in the era of the spaceship Enterprise (from Star Trek) – where we ask the onboard computer: “When did we last cross the speed of Warp 6.7 on the way to the Gamma Quadrant?”. And the instant cheerful reply: “Last Tuesday at 3:32 pm”.

Well, cognitive access to complex industry information is a lot closer than we think. In fact it’s here. How? For a machine to understand natural language-based questions fully it must break down the syntax of the sentence and understand the meaning and context of all words it encounters. Some words and phrases are common knowledge, such as “when”, “speed”, “on the way”. But others are industry specific, such as “Warp”, Gamma Quadrant”. And then there is the matter of resolving the sense in which words with dual meanings are interpreted in the context of the sentence, such as “cross” which is a verb, and also has a couple of meanings as a noun. And notice the meaning of the word “last” in both the question and the answer? Confusing, you may say ?

A definite answer lies in the field of knowledge-based natural language understanding (NLU).  Building and leveraging knowledge that is specific to the commonly understood world and to specific industries helps a computer fully understand the question. It’s not just intent that is gleaned but the entire question is understood in every detail. The computer then locates and provides the required information with high accuracy.