Full Nelson: Recommind's Enterprise Search On TechWebTV

Groupware. Portals. Enterprise search. I'm not saying they're irrelevant, I just sort of forgot about them. Like tricycles, ER, and Oasis; the use of the word "bashful." But they're all relevant in some way (except probably Oasis), especially enterprise search: Witness -- speaking of bashful -- Microsoft's recent purchase of Fast (see video below for a fun perspective from Steve Ballmer at Web 2.0 on Microsoft and search).

And it turns out that the market is growing at 39%, according to this article. It's especially meaningful at a time when regulations about document retention are omnipresent. So it's little surprise that Recommind has made (made up?) a name for itself in conducting searches across all types of disparate data within a corporation. (See video interview below.)

With the exploding growth and evolving types of information, especially those coming from newer Web applications, search tools must get more sophisticated and, frankly, more relevant. The drivers include the usual scary things that emanate from "the man": government regulation, the threat of lawsuits, the fear of losing knowledge. As you might imagine, financial and health care companies are on the top of Recommind's target (and current) list of customers.

Recommind's MindServer technology looks at the language within documents, and the company says it can extract concepts and correlate them with concepts in other documents. In fact, it's created its own marketing lingo for this (brace yourself): conceptual search platform. And it works in any type of document that has a contextual nature. Let's review: not just content, not just context, but also concepts. Its search technology is based on statistic modeling it calls probabilistic latent semantic analysis. Sounds sexy.

Recommind also has made its presence felt in the e-discovery arena (if you don't know what this is, it's a good thing; in fact, I would highly recommend e-avoidance, but that's for another day). It has been part of a working group that has created an XML schema called EDRM, which allows sharing of information among repositories for the specific purpose of providing information during lawsuits.