An abridged version of this paper appeared in R. Ennals and P. Molyneux (eds.), Managing with Advanced Information Technology. London:Springer-Verlag. 1993

1. Why is Artificial Intelligence important to business? The 'White-Collar Overhead'

Not so long ago -- date it, if you like, to the appearance on the Prague stage, in 1921, of the very first robots in Karel Capek's visionary play, RUR -- the project of introducing 'intelligent systems' (or however else we might want to designate them) into the industrial context would have been equated specifically and uniquely with the replacement of blue-collar labour by some form of unwearyingly complaisant 'mechanical man'. Automate your factory with machines that will cheerfully and tirelessly grab and move and shape and weld things, and—Toshiba through Toyota—you will produce high quality goods in bulk at low cost.

Or at least in theory. But the costs of initially 'robotizing' the factory are high, while at the same time the increase in productivity and cost saving has turned out to be disappointingly low. Studies over the past decade have clearly shown why this is: blue-collar work in the manufacturing process accounts for possibly as little as 25 to 30 per cent of costs, such that, even if factory automation were cost-free and all blue-collar work could be done by robots, this would represent a saving of only around one third of a company's production costs.

At the same time, it has been estimated that for as little as perhaps only 2 per cent of the time an order is in a factory are value-added operations being performed, and that white-collar labour may account for as much as 50 to 75 per cent of production costs in the manufacturing cycle (report published in American Machinist, Vol.124). The bulk of the costs, that is, are consumed by white-collar knowledge-based tasks such as procurement, design, quality assurance, management co-ordination and control, marketing, sales and distribution, service, and financial management. If the productivity of personnel involved in knowledge-intensive work—in planning, problem-solving and decision-making—can be enhanced, then costs fall significantly, and the company consequently becomes more competitive.

The potential business benefits of Artificial Intelligence (AI) and knowledge-based systems (KBS) in this context are inestimable: a new generation of commercial AI systems promises to give economic leverage to companies through their ability to find fast and often novel solutions to complex problems that are either time-consuming for human beings, or hard to solve, or maybe even not solvable with alternative, more traditional, computing technologies. Already many hundreds of organisations, large and small, and including such names as American Express, British Telecom, Coopers and Lybrand, and the New York Stock Exchange, are successfully using AI systems for a wide range of knowledge-based tasks—from the interpretation of international tax laws, through fault diagnosis, identifying ships from satellite pictures, mineral prospecting, and machine translation, to the allocation of landing gates at busy airports, and much more besides.

The principal aim of AI, in short, is to endow machines, not with shop-floor muscle, but with the ability to perform tasks requiring the application of types of general intelligence and often specialist knowledge that one associates uniquely with the human mind. Although the last decade has brought unprecedented levels of funding, interest, and concomitant expansion to the field such that there is now probably no definition of the discipline that matches more than a subset of the contents of contemporary AI journals, conferences, and courses, the classic definition offered by Marvin Minsky has probably stood the test of time better than many:

artificial intelligence is the science of making machines do things that would require intelligence if done by men. (Minsky, 1968)

This chapter will first add flesh to the definition by outlining in a little more detail what AI is; in the remainder of the chapter I will briefly survey some of the applications of artificial intelligence and knowledge-based systems in the white-collar sector.