Jeeves One Enterprise Flexes its Tentacles

There's no question about it: customer service on corporate web sites seems to be improving by leaps and bounds, and it's due in large part to navigation technology companies like Ask Jeeves. Jeeves has for some time been providing enterprise solutions for larger companies to help them deal with common customer inquiries using natural language. Its product has made some considerable strides recently, and it's worth taking a closer look.

Inspired in part by the buzz created by books like Seth Godin's Survival is Not Enough, companies have been paying more attention of late to the theme of constant improvement and adjustment by making better use of market research data. (I just stumbled across a newspaper ad with the title "CIO + 683% ROI = CEO." I doubt anyone even knew what a CIO was five years ago; now they're the subject of ads in the business section.)

There are few better ways of finding out what customers really want from you than analyzing the thousands of questions they type into a customer service web site's search box every day. That's part of the service offered by Jeeves One Enterprise: analytics of customer inquiries.  According to James Speers, product manager for Jeeves Solutions , the enterprise software division of Ask Jeeves, "customers' unconstrained online questions offer more insight about market demand than traditional metrics like click-path analysis and the like." Nike, for example, discovered that many customers were searching for a hiking-type shoe with more support. In other cases, it may be a case of a large number of customers not able to access a particular piece of information because it isn't easy to find. Some companies receive upwards of 20,000 questions a day. That's a lot of feedback.

Analysis of data is one thing, but how well does Jeeves do in serving up relevant info to consumer queries? Doing so is no small challenge; as Delphi Group's Hadley Reynolds points out, "the answer the customer is looking for usually depends on data from a variety of systems scattered around the firm." The problem, points out Speers, stems from the fact that for many large companies, the "web buildup was deployed along organizational lines," rather than being tailored to customers' primary informational needs.

Jeeves' recent acquisition of Octopus, an early entrant into the "metabrowser" space, has given it some of the fundamental technologies that can help make it possible to pull this disparate data together. The recently-released Jeeves One Enterprise comes with this deep data aggregation capability and includes customized "knowledge packs" which allow companies to start providing sensible answers to common queries based on what Speers calls a "giant ontology and lexicon" and "question templates based on Ask Jeeves' collected wisdom [historical data about customer inquiries in different industries on both the consumer ask.com search engine and corporate web sites]." Pricing starts in the low six figures.

Jeeves is already providing natural language web self-service functionality to dozens of large corporate clients, including Visa, Nike, Ford, Nestle, and Verizon Wireless, as well as government clients like the State of Washington, all of whom should welcome the more agile tentacles of the Octopus-enhanced Jeeves One Enterprise.

To my mind, the most attractive aspect of natural language navigation technology and its attendant analytic capability is the fact that it sets up a dynamic dialogue with a company's market. As Godin argues (Survival is Not Enough , page 88), the company that stays in business is often not the one whose CEO clings to "spreadsheets and analyst reports and the macho certainty that their vision of the future is correct," but rather one that admits "we're not sure what the future brings, but we are sure that we've got the fast feedback loops and bias to evolve that we'll need to stay ahead of the competition." If thousands of your customers are typing questions into the search box on your company web site every day, shouldn't you be listening?

Andrew Goodman is Editor of Traffick.com.