Intelligence meets the network: agents show glimmer of futuristic Net Promises – intelligent agents

Intelligence meets the network: agents show glimmer of futuristic Net Promises – intelligent agents – Internet Edge

Guy Sandra

Imagine having a virtual agent–a technological genie in a bottle-ready and willing to fulfill your commands. It knows your every whim and makes it so.

To schedule an airline flight, simply ask the software agent to find the cheapest ticket. Based on your seating, eating and other preferences that the agent already knows, it negotiates the best deal, schedules the flight and updates your personal planner.

Sound too good to be true? It shouldn’t because intelligence has come to the network.

Intelligent agents go a step beyond push technology by finding only the information end users want, rather than sending gobs of data that they must browse through to uncover what they need.

Although no universal definition of a software agent exists, agents generally perform tasks that users explicitly approve. Users can file personal profiles in two ways–actively, by spelling out likes, dislikes and interests, or passively, by using software that intelligently “watches” users’ habits, including the World Wide Web sites they visit.

From there, the agent software can take its best guess, says Paul Grand, founder and chairman of NetCount, a Los Angeles company that tracks Web site and on-line advertising performance.

Service providers on the cutting edge

Carriers are using intelligent agents to help their employees and customers stay ahead of today’s information-driven game.

US West !nterprise Networking Services helps its business users find federal government contracts with its !nteract Procurement Services (!PS) solution. !PS uses transactions based on electronic data interchange to let the businesses bid on government requests for quotations and proposals, or to place orders and ship and process payments over the Internet (Figure 1).


!PS’ Federal Procurement Services application continually scans and selects government proposals based on each user’s profile. It instantly notifies users by e-mail or pager when a match is found.

“That saves companies lots of time and money,” says Miles Morimoto, director of !nteract Transaction Services. Otherwise, businesses must rely on employees to look through newspapers, magazines, commerce journals and business dailies to find such opportunities, he says.

Once the match is found, the company need only insert a price into its bid rather than retype its proposal, says Michael Sabo, director of !nterprise’s Internet Services Development (Figure 2).


One of the system’s users says it is a must-have in today’s business world, especially because the federal government requires purchases between $2500 and $100,000 to be done electronically. But she is concerned about the loss of face-to-face contact.

“It’s a gamble for us to do it this way because we’re not working elbow to elbow with people. We can’t say, `Yes, I see your situation, and I’ll show you what I have that might solve the problem,'” says Claudia Holtz, business administrator for Mitchell Lewis & Staver Co., a Wilsonville, Ore., wholesale distributor of water systems, agricultural sprayers and compressed air equipment.

But she is excited about U S West’s plans to expand the program to business-to-business e-commerce. “It’s another way to reach customers,” she says.

This type of matchmaking is called collaborative filtering, and its biggest proponent is Firefly Network, a Cambridge, Mass., developer of open profile management and advanced personalization technologies.

Barnes & Noble is using Firefly’s Catalog Navigator 2.0 to match readers with peers and books they will find interesting and useful. The technology uses readers’ ratings of articles and authors to find similar works among a huge catalog of books.

“It’s automating the word-of-mouth process,” says Liesel Pollvogt, communications manager for Firefly. “If I’m looking for a new book to read, I go to certain people because I know we have similar tastes. This allows that process to happen automatically on-line with a huge community of people from throughout the world.”

MCI uses intelligent agents to help executives create their own profiles. The agents also help networkMCI librarians monitor and pass along critical information to those executives. “The [news] stories coming in number 8000 to 12,000 a day. There’s no way we can go through that manually,” says Diep Truong, manager of networkMCI library’s application services.

The carrier also uses intelligent agents to filter industry news and make it available to employees on the networkMCI library Web site. The job takes one-fifth the time it once did, Truong says.

“The software just gives you the foundation. The last piece of it–refining the query and designing how the information comes together–we think is very important,” she says.

Filtering unwanted messages

Intelligent agents are sure to get smarter in the future. They will be used in any number of appliances and across different operating system platforms for stationary and mobile users, says Doug Ehrenreich, who left Sun Microsystems in August for Softwire, a startup middle tier solution developer in Larkspur, Calif.

This increased adaptivity will help carriers build out some of the new services and applications that users are demanding. “Intelligent agents will start opening up a whole new form of intermediation, affinity groups or vanity services on the Internet because they enable real one-to-one marketing and very customized service environments,” Ehrenreich says.

Carriers likely will take the technology a step further by building intelligent agent software to predict market trends, especially as competition heats up, he says.

Agents will empower end users as well. People who dislike being bothered by telemarketers at dinner could program their future “smart phones” to block a list of predetermined numbers, says NetCount’s Grand. “Your phone will know it will not take calls from a certain list of numbers. Before the phone even rings, as the call comes in, it will check the list.”

In case anyone is reminded of George Orwell’s “1984,” intelligent agents do have a dark side. When a user seeks information, his or her preferences can be directed to a server. And those preferences can be used by advertisers.

For example, an advertiser can find out when a user’s car lease expires, review the user’s habits on surfing car-related Web sites, and, four weeks before the lease expires, send a targeted message about a car sale to the user.

“Is it really right for a company or anyone to be shaping people’s desires?” Grand says. “It’s direct marketing to the extreme.”

Concerns similar to those associated with the V-chip also are being voiced. What happens to society when people receive only news and information that they want to see–effectively censoring themselves?

Ehrenreich sees no need to fear Big Brother. He compares the situation with the hundreds of magazines that fill new-stand racks. “When I want to dig deep into an environment, that’s when I’m going to set the parameters to really go out and pull the information,” he says.

Grand believes each individual must be as independent as the Internet itself. That means getting empowered with information rather than letting the government, businesses or organizations dictate the content flow.

COPYRIGHT 1997 PRIMEDIA Business Magazines & Media Inc. All rights reserved.

COPYRIGHT 2004 Gale Group