previous | contents | next

232 Marketing and Sales

planned. It is practically impossible to construct a model that can tell how long it will take, or how expensive it will be, to develop an emerging market.

Several high-tech ventures started up in the mid-1980s using rule-based systems technology developed by computer science's artificial intelligence (Al) community. These firms have evolved rather slowly and together have annual sales of less than $50 million. The technology-to-product transition has been slow to occur, with a relatively small number of applications using the rule-based programming approach. The organizations that have remained small and operated in a controlled fashion have been successful.


The first Al companies. Teknowledge was founded in 1981, along with three other firms-Al Corp., Intellicorp, and the Carnegie Group (based on Carnegie-Mellon's work in Al). Teknowledge's mission was to become known as the premier Al company. Its strategy was to hire all the best people so that no one else could start up. It rented expensive space in Palo Alto, hiring Al researchers from Stanford's Al laboratory who had little or no product experience.

Teknowledge made "strategic alliances" with several large U.S. and European companies, including GM, by selling stock in exchange for a close working relationship on applications. It initially trained other organizations in rule-based programming, built special systems, and did government research in order to develop its next-generation technology. All three of these activities were potentially profitable, and the firm could have run profitably from the outset. Since such a venture was service- and labor-intense, Teknowledge attempted to build a high-standard product in order to obtain higher operating margins and higher valuation.

New, small, low-overhead competitors, such as Neuron Data, began to introduce standard rule-based products for a small fraction of Teknowledge's price. They had small staffs and equally bright people, but unlike Teknowledge, their people had product development and start-up experience.

Teknowledge raised tens of millions of dollars through private and public funding. In 1988, with money still in the bank and sales of its expensive product rapidly declining, Teknowledge merged with ailing Cimflex, a Pittsburgh-based company building custom programs for manufacturing. The Teknowledge portion of the organization was reduced to a small operating division doing contract research. Those at Teknowledge who were responsible for the merger (and who were also founding shareholders) did not suffer financially.

In contrast, Intellicorp (which began as IntelliGenetics) started from the same Stanford core technology to do gene sequencing. After the firm went public, the gene-sequencing business turned out to be less robust than it had originally thought. Intellicorp went on to develop and market its proprietary language, KEE, for building rule-based systems.

previous | contents | next