Back to Contents of Issue: October 2000
by Daniel Scuka
Silver Egg Technology claims its AIgent gets to know you well enough to make shopping recommendations that aren't inane.
Neither Einstein nor his still-peripatetic brain ever made it to Japan,* but if he had, he'd appreciate the artificial intelligence (AI) being built into Silver Egg Technology's one-to-one Web marketing engine, dubbed AIgent by the firm's 38-year-old CEO and co-founder, Tom Foley. "This technology is called a recommendation engine, and in its simplest form, it observes individual customer behavior. It looks at clicks, purchases, or survey data, and combines all available information in real-time to build a customer profile. Based on that, it adapts and targets the content that the user receives," explains Foley.
While it may sound like rocket science, true one-to-one interactivity is a Holy Grail that has a lot of Web vendors tremendously excited. (Granted, it's not really "one-to-one" if a bot is one of the ones, but ...) Banc of America securities analyst Greg Vogel says, "There is a growing demand for the ability to interact directly with customers over the Internet, and provide sales and marketing, customer support, and client and customer relationship management."
But that takes a lot of time and is very expensive. If Web vendors can get a software company to provide that capability for them, that's a much easier route. And Silver Egg may be ideally positioned to catch the first wave of any Net vending stampede in Japan, where the one-to-one marketing sector looks set to take off.
Michiko Kurata, marketing manager for Silver Egg competitor BroadVision Japan, reports they expect to see 200 percent revenue growth in 2000.
"This year, Japanese companies understand that Web businesses need special models, and they're moving from using the Web as a strategic marketing tool to deploying 'smart' one-to-one methodologies," says Kurata.
And smarts is one thing that Silver Egg's AIgent doesn't lack. The system employs a synthesis of three different AI technologies: knowledge representation, machine learning, and matching. How does this work in practice? "Our technology is based on the idea of deep knowledge," explains CEO Foley. "What enables a good salesperson to give great service is that he or she knows their products well, and can match them to the customer."
Foley offers the analogy of purchasing a CD at a bricks and mortar record shop. A good sales rep might recognize a customer who bought a Chet Baker album last month. If the rep knows his jazz, he would recall that Chet Baker was a trumpet player of the "cool jazz" genre from the late '50s and infer that the customer likes cool jazz. He could then draw on his jazz knowledge to recommend a CD likely to pique the customer's interest from Miles Davis' cool jazz period -- and induce a fresh sale.
Another key aspect of AIgent is customer profiling, where one customer who clicks on a particular link may be served up content different than another who clicks on the same link. The server uses the customer's responses to pop-up questions and browsing history to build a profile.
"This is like what Amazon does when it recommends books to purchase, but with AIgent the synthesis of data is more complex," claims Foley.
Foley's bullish on the market potential here. "We expect to do over $1 million in business this year. We're concentrating on a small number of showcase sites that will feature our technology and allow us to build up some market history. We already have five clients, and more are coming on board."
Potential clients include anyone trying to offer a service via the Web -- think CD or video sales, computer games, books, software, travel and ticketing, job matching, and real estate. Silver Egg offers a version of AIgent suitable for new and medium sites, and from October the product will be available through resellers and VARs.
Kevin Williams, a research manager at IDC Japan, agrees that artificial intelligence may be the marketing technology of the future. "The options are endless," he says, "but implementation of such technology really needs to be carefully analyzed." Williams warns that the technology is not without its risk -- especially where customer acceptance is concerned. "Offline, customer profiling in the United States has received widespread criticism in the past. Consumers feel that their freedom to choose products and services themselves is greatly diminished. But that is not to say that profiling is useless; on the contrary, if it is done in a non-intrusive, non-insulting way, it could revolutionize many industries. AIgent, or similar services, are necessary given the dearth of customer management seen in Japanese SMEs," he adds.
Why did Foley pick Japan? "All the action is here," he answers. "That's why all the VCs are here. There are lots of fresh opportunities in this country." Silver Egg is now looking for second-round investors, and if AIgent's technology works as promised, future investors should be at least as pleased as future customers.
* Note: for more on Einstein's wandering brain, access http://salon.com/books/review/2000/07/06/paterniti/print.html
Note: The function "email this page" is currently not supported for this page.