In a recent Businessweek article, the writer pointed out that IBM’s CEO promised that Watson, their supercomputer that had success answering tough trivia questions on the TV show Jeopardy, will generate $10 billion in annual revenue within 10 years. This is surprising, given that the CEO also noted that Watson is currently generating business at a rate of about $100 million per year.
Businessweek also pointed out that according to a review of internal documents and interviews with Watson’s first customers, Watson is having trouble solving real-life problems versus Jeopardy-type trivia questions. For example, just a few months earlier one of Watson’s first projects with the M D Anderson Cancer Center was described by the IBM executive overseeing the Watson business as being “in the ditch.” A huge challenge with these real world projects is the fact that Watson’s basic learning process requires IBM engineers to master the technicalities of a customer’s business – and translate those requirements into usable software. That process is very challenging, given the subtle complexities in most businesses.
Given that the technology that Watson is based on is the arcane field of artificial intelligence(AI) that has a long history of being very difficult to put to work in a practical setting, it didn’t help IBM’s credibility to hear the assessment Klaus-Peter Adlassnig, a computer scientist at the Medical University of Vienna and the editor-in-chief of the technical journal Artificial Intelligence, who said that Watson can only acquire “flat and broad” knowledge from medical texts and case studies, and hence in a clinical setting, the computer would make for a very thorough but crippling literal-minded doctor – not necessarily the most valuable addition to the medical staff.
This IBM example provides a couple of good reminders:
1.) Beware of Unsupported Promises – Citing wild projections when you have absolutely no supporting data is dangerous. People understand immediately that you are flying blind and pretending like you know something you don’t. Given the difficulties that people have had in successfully applying AI to real world problems, you know that IBM’s CEO is pulling figures from the air.
2.) Build Credibility with Results – Gaining trust and confidence is all about generating a sequence of positive results. What will eventually sell Watson’s services will be very impressive examples of Watson making big contributions to some very tough real-world problems. IBM should wait for such examples before it makes such big claims about future potential.
These two points apply not only to organizations but also to individuals.