The AI That Saved a Company $40 Million a Year
Digital Equipment Corporation's R1/XCON became the first expert system to prove AI could deliver massive commercial value, configuring minicomputers faster and more accurately than any human.
Alex P. Kok / CC BY-SA 4.0
R1 currently saves DEC more money annually than it cost to build and maintain over its entire lifetime.
— John P. McDermott
In 1986, the R1 (XCON) expert system, developed by John P. McDermott and later maintained by Virginia E. Barker, revolutionized the way Digital Equipment Corporation (DEC) configured and sold its VAX computer systems, saving the company an estimated $40 million annually.
What happened: In 1978, John P. McDermott created the R1 (XCON) program, a production-rule-based system written in OPS5, to assist DEC in the ordering of VAX computer systems. By automating the selection of system components based on customer requirements, R1 significantly streamlined the configuration process. Virginia E. Barker played a crucial role in maintaining and improving R1, contributing to its commercial success. At its peak in 1986, R1 processed 80,000 orders per year, becoming the most commercially successful expert system of the decade. Xcon - Wikipedia
Why it matters: The success of R1/XCON validated the potential of expert systems in the business world, leading to a surge in development and investment. However, its eventual maintenance challenges also highlighted the limitations of early AI systems, foreshadowing the expert systems crash that would follow. R1 (expert system) and Digital Equipment Corporation
Further reading:
Why This Mattered
R1/XCON was the most commercially successful expert system of the 1980s, validating the entire expert systems industry. At its peak it processed 80,000 orders per year and saved DEC an estimated $40 million annually, triggering a gold rush as Fortune 500 companies scrambled to build their own. Its eventual maintenance nightmare also foreshadowed the expert systems crash.
