Tech Success: Autonomy makes job hunting easier at Labor

IT solutions in action

Project: Advanced search capabilities

Agency: The Labor Department's Employment and Training Administration

Partners: Autonomy Corp. plc, Cambridge, U.K., and Information Strategies Inc., Washington

Goal: The Employment and Training Administration wanted to offer advanced search capabilities for both its constituents and its employees.

Obstacle: The agency houses more than 1.2 million documents. Although some of them are Web pages and easily searchable, others are in formats that are more difficult to index, such as PowerPoint.

Solution: Autonomy's search solution can index documents in more than 300 formats, so it can provide the necessary tools for searching through multiple types of files.

Payoff: The public has more advanced capabilities to find documents by searching rather than by navigating through multiple directories. Employees can cut the number of searches they perform in a day by using an integrated search engine to query several formats at once.

Information Strategies taps software firm for robust search engine

When job seekers and employers come to the Labor Department's employment and training Web site, they need to do more than just search HTML pages. They need to search through documents that are in Word, PowerPoint or portable document formats.

To meet the challenge of building a search engine that could comb through its vast storehouses of data, the agency picked software by Autonomy Corp. plc of Cambridge, U.K., and hired Information Strategies Inc. to install it.

"Autonomy offers a powerful infrastructure for dealing with unstructured data," said James Townsend, president of Information Strategies, a Washington-based knowledge management solutions provider. Unstructured data are text documents, e-mail and other documents that aren't contained in a database.

For Townsend, Autonomy simplifies sifting through these documents by providing an integrated search system.

"You search your e-mail one way, you search the Web another way, and you search your files another way. If you can do one search that pulls from all these sources, then you can improve productivity," Townsend said.

In January, Information Strategies won a contract to implement Autonomy's solution within the Labor Department's Employment and Training Administration, which runs the Web site. The administration's mission is to help workers find training and employment. It also gives businesses a conduit to find potential workers.

The administration chose Autonomy because it wanted to offer a search service for both its users and employees that could respond to more exacting queries. Not only did the agency want the capability to search through Web pages and forms in the portable document format, but files in Word and PowerPoint as well.

"They wanted something that was much more advanced than a full-text search," Townsend said.

To implement this solution, Information Strategies used Autonomy's Classification Server software with a number of plug-in modules to facilitate searching in particular formats. Autonomy's software can search through 300 different document formats.

Given a search term, the Autonomy software returns not only a list of documents, but also documents that may be related conceptually to that term. Autonomy also permits searches to be refined. If the first query doesn't return the correct answer, the user can modify the search term.

Information Strategies completed the installation in June and now is supporting the software through a maintenance contract. The search engine for public information can be used at the administration's Web site: http://www.doleta.gov

Autonomy officials see the potential for expanding the company's federal presence beyond the Labor Department.

"Almost every agency has problem with unstructured data. It is easy to think of agencies that do analysis that depend on this kind of pattern identification. The intelligence community, the Department of Defense, the medical sector of the government," Townsend said.

In the competitive market of search engine technologies, Autonomy vies against companies such as Convera Corp., Vienna, Va., and Verity Inc., Sunnyvale, Calif.

"We differentiate ourselves from traditional search engines in the high degree of automation we bring to the table. Autonomy helps the users to completely bypass the need to search for information," said John Cronin, vice president of the government sector for Autonomy. The software can deliver documents to end users even before they request them based on other documents they open.

For 2002, Autonomy reported revenue of $51 million and income of $6.1 million. The company recently completed its purchase of Virage Inc., which developed software for searching through video and image files. With this acquisition, Autonomy can better compete with Convera, a company whose specialty is video searches, Cronin said.

Cronin said the cost of setting up basic Autonomy capability would run about $100,000, and a typical federal solution supporting a few hundred users would run about $350,000.

The Energy Department, the Joint Chiefs of Staff, the Securities and Exchange Commission and various intelligence agencies use Autonomy's search solution. The National Park Service is trying out the software in a pilot project as well.

In the government market, Autonomy does not sell direct, instead relying solely on integrators and resellers, Cronin said. It has about 20 federal partners, including Accenture Ltd., the government solutions unit of AT&T Corp., General Dynamics Corp., Raytheon Co. and Science Applications International Corp. In the reseller space, the company works with GTSI Corp. and IGov.com.

Autonomy's product is infrastructure software, meaning that it sits between user applications and data, Cronin said. Although the company offers a portal interface, the software is more often used in conjunction with third-party applications, or it is inserted behind custom-written software.

"Our model in the federal market place is to rely on proven systems integration firms to ensure that systems are properly integrated," Cronin said.

Joab Jackson, Government Computer News associate editor, can be reached at jjackson@postnewsweektech.com.