Enterprise Will Value Natural Language Search

Don’t bother remembering the name Powerset because if they build a natural language search (NLS) engine that actually understands the concepts of time, place and sentiment, Google or Microsoft will snap them up.

Danny Sullivan is skeptical about Powerset’s prospects — too many broken NLS promises in his past. He points out that most of us use two keywords in our searches. Heck, I find it amazing that a lot of folks search for “Google” from their Yahoo toolbars… If that’s the way we want to search, why do we need a natural language search engine? Esther Dyson is more positive (and an investor).

I’m hopeful that Powerset — or another company — cracks the code because I believe NLS will be a powerful application for enterprise. NLS could solve tough problems we are dealing with right now. For instance, my clients want to find their customers who might be having issues with the products they’ve purchased so their customer service reps can solve their problems quickly to keep the customer and generate positive WOM. They want to gather competitive intelligence on-line to expose opponents’ vulnerabilities and to entice competitors’ disappointed customers.

Today, monitoring for sentiment is a major manual effort. And, as we all know, manual processes don’t scale easily. NLS could make the enterprise search process more efficient and effective.

Don’t get overly excited about NLS today, however. I suspect we’re a year or more away from having viable NLS solutions for enterprise.

5 October 2006 | Monitoring, Search | Comments