Previous | Table of Contents | Next |
Some examples of agents include those for filtering E-mail, scheduling appointments, locating information, making travel arrangements, and paying bills. It has been suggested that agents will someday perform the tasks of a knowledge worker throughout the day, making human executive assistants unnecessary.
The World Wide Web has five primary shortcomings:
The increasing use of agents on the Web has resulted in a growing number of creative and helpful agent solutions that address these shortcomings. The following sections present five types of Web agents search engine, monitor, publisher, guide, and personal assistant each of which addresses a major Web shortcoming and supports several management needs.
The ease of creating Web pages has triggered the proliferation of Web pages representing users, organizations, and topical issues. Burgeoning numbers of new companies providing skills and services promote further production of Web pages. Unfortunately, the proliferation of data available on the Web makes it difficult for managers to find useful information.
To address this problem, intelligent agents called search engines seek information and present results based on prespecified criteria. Search engines extract, filter, and compress critical data from a broad range of internal and external sources. Most search engines have similar interfaces in which a manager enters some criteria (e.g., subject = stock market AND date = 1996). After the search is invoked, a list of Web page addresses are displayed for further investigation. Some search engines provide options that set a time limit for the search, the number of addresses to be displayed at one time, and the amount of detail to include in the output.
One search engine, SavvySearch, is considered a metasearch tool that provides a convenient interface, available in multiple languages, to several other search engines.1 The manager enters criteria for a search and selects options that expand or limit the search output. This metasearcher ranks a list of available search engines (e.g., Alta Vista, Lycos, and Yahoo)2 and transfers the search request to selected engines. The selection is based on the query, topic area (e.g., Web resources, news, or entertainment), estimated Web traffic, anticipated response time of the other search engines, and the load on the SavvySearch computer.
Not only is there an abundance of information on the Web, the information is changing constantly. Many Web pages contain information that is regularly added to, deleted, or modified. The location of these Web pages changes as well; servers, directories, and file names move and disappear, frustrating managers who reference invalid links.
Monitor agents accommodate the problems created by dynamic information. These agents let managers track critical data by looking for changes on the Web and communicating these changes to the manager. Instead of relying on people to determine when information becomes obsolete, a manager can count on monitor agents to flush out changes as information evolves and provide direct notification of them.
Both Specter Communications and First Floor Software offer agents that search the Web for updates on selected sites.3 WebWatch from Specter Communications, for example, checks selected sites automatically and highlights modifications. A manager can have WebWatch monitor competitor sites for market changes, for instance. Managers who use WebWatch can view downloaded Web sites offline at their convenience. WebWatch has filtering capabilities similar to search engines that allow the agent to traverse the Web looking for changed Web sites, while the program resides on a personal workstation and updates Web bookmarks(i.e., listings of Web site addresses). Such an agent helps managers keep pace with the ever-changing business world.
Information on the Web is unordered, redundant, and uncategorized. Unlike a library, which offers a well-defined process for searching for information, the Web offers no standard approach for meeting information needs. Therefore, intelligent agents that custom tailor information to individual executives while accessing and integrating a broad range of internal and external data are highly desirable.
Publisher agents do just that. Although they have access to large amounts of information, they present to a manager only the topics and types of information that have been prespecified, often in a variety of presentation formats.
For example, PointCast Network (PCN) provides a personalized newspaper to users of the software.4 First, managers select the topics that they want to read about. This could include headline news, stock quotes, weather, sports, or industry and company news. Based on a managers interests, hundreds of articles are filtered, integrated, and pointcasted to the manager to read at his or her convenience. The text is accompanied by weather and stock price quotes presented graphically in maps and charts.
Previous | Table of Contents | Next |