Search engines and intelligent agents have only just begun to return their full business value
 

Seek, and Ye Shall Find

The Web is the richest repository of information in human history, but few companies take full advantage of it. Although they may well consider the Web a valuable source of business intelligence, their interactions with it tend to be manual, haphazard, and isolated from mainstream BI processes. Instead, most companies focus on the Web’s utility as an outbound channel for communicating with customers, as a platform for transactions, and, only recently, as a medium for collaborating with suppliers and other supply-chain partners.

All well and good, but as Dan Sullivan explains in “Eye on the Competition”, the Web also presents an opportunity to automatically gather “competitive intelligence” (or “market intelligence,” as some would call it) about your business environment, whether the sources contain public disclosures about your competitors’ activities, general information about regulatory developments or new technologies, or even newsgroup discussions among your customers that concern their perception of your organization.

In Sullivan’s document-centric approach, search engines and Web robots dynamically gather the grist for the mill, but text processing and mining tools, many of which still require manual intervention, prepare that competitive intelligence in a document data warehouse for general consumption. In the more expansive “Web farming” concept endorsed by Bolder Technology’s Richard Hackathorn, a general-purpose data warehouse is the main information artery into which intelligence extracted from the Web flows, joining data derived from internal sources. In either case, a competitive intelligence infrastructure becomes “the eyes and ears of the enterprise,” to use Hackathorn’s words.

The growing prevalence of extensible markup language (XML)-based content on the Web will facilitate the competitive intelligence process. But more interesting is the role that intelligent agent technology potentially has to play—perhaps as the means by which text transformation and mining can occur automatically, and if need be, results delivered to your desktop.

The Sentinel

Agent technology is hardly new. Enterprise systems management (ESM) platforms such as BMC Software Inc.’s Patrol and Computer Associates International Inc.’s Unicenter TNG heavily leverage such technology in their respective architectures; in those products, agents continuously monitor and report on the health of enterprise systems in an event-handling process. But according to Pricewaterhouse-Coopers ESM consultants Elmar Hussman, Thomas Schmitt, and Thomas Schuler (“Agents Are Watching,” page 38), those systems only scratch the surface of what agent technology can do. In their view, classical ESM systems fail to fulfill the new requirement for a comprehensive, business-oriented view of the enterprise, but their event-handling approach serves as inspiration for a true “e-business process monitoring” architecture.

In such an environment, which currently exists as a PricewaterhouseCoopers prototype, information about business process execution and events—as well as the status of IT systems—is dynamically gathered by intelligent “cells,” each of which represents a logical component, a business process, or IT infrastructure. These cells, which are layered in a rule-based hierarchy, have specific event-handling properties and log the events in which they are “involved.” Hence, by “questioning” these cells, as the authors put it, you can easily deconstruct business-process failures to their technical sources. For example, in this environment, you could trace a business event such as the failure of an online order to a problem in the ERP system, all through a single e-business monitoring interface. Now that’s business intelligence.

Cat People