IBM Wants to Put a Watson in Your Pocket

The Jeopardy! computer could answer health and finance queries
Watson’s IBM servers can analyze documents, websites, and books at a rate of 66 million pages a second

Illustration by Peter Arkle

IBM researchers spent four years developing Watson, the computer smart enough to beat the champions of the quiz show Jeopardy! Now they’re trying to figure out how to get that brainpower into your pocket.

Bernie Meyerson, IBM’s vice president of innovation, envisions phones tapping into Watson, which would answer complex questions remotely, like a supercharged version of the Siri personal assistant built into Apple’s iPhone. A farmer, Meyerson says, could stand in a field and ask his phone, “When should I plant my corn?” He’d get a reply in seconds, based on location data, historical trends, and scientific studies.

Finding additional uses for Watson is part of IBM’s plan to tap new markets and boost revenue from business data analysis to $16 billion by 2015. After mastering history and pop culture for its Jeopardy! appearance, the system is analyzing Citigroup clients’ financial needs and crunching cancer data for WellPoint. Watson 2.0 would be energy-efficient enough to work on smartphones and tablets. “One day, you will have ready access to an incredible engine with a world knowledge base,” Meyerson says.

Although the new service doesn’t yet have a release date, Watson is already the company’s most high-profile product since the personal-computer unit was sold to Lenovo Group seven years ago. IBM will need to figure out how to price and deliver Watson for different corporate customers, likely on a case-by-case basis. First, though, the company will have to overcome the technical hurdles of making it handheld. Watson’s nerve center is 10 racks of IBM Power750 servers running in Yorktown Heights, N.Y., with the processing power of 6,000 desktop computers. In its original form, the system consumed 80,000 watts an hour answering questions; even though most of Watson 2.0’s computations would occur at the data center, a widely used smartphone app would require too much power for it to be practical.

Illustration by Peter Arkle

Watson also takes awhile to do the “learning” necessary to become a reliable assistant in specialized subjects. The Watson deal with WellPoint was announced last September, and the system won’t master the field of oncology until at least late 2013. To prepare Watson for its Jeopardy! appearance, IBM researchers had the machine scour documents, websites, and books stored on its servers—which it can analyze at 66 million pages a second—so it could answer queries based on how similar questions had been addressed before. After it’s been told enough times which answers are accurate, Watson eventually develops an expertise.

IBM researchers also plan to add image-recognition capability so Watson can respond to real-world input, such as an X-ray, says Katharine Frase, the company’s vice president of industries research. “In 2.0, we hope to give him more senses,” she says. “A guy could say into his phone, ‘Here’s where I am, and here’s what I see,’ lifting it up to take in images of the environment.”

Unlike Apple’s approach with Siri, IBM is focused on corporate customers and on tackling complex questions. “Taking this deeper approach with Watson, having it specialize in business areas, could be deeper in value and deeper in revenue than a Siri,” says Andrew Bartels, an analyst with Forrester Research. “It matches how IBM approaches markets, and it makes business sense to go mobile.”

Illustration by Peter Arkle

As the technology improves, a mobile Watson could be an extension of services that IBM already offers to business customers such as WellPoint, the second-biggest U.S. health insurer, says Martin Kohn, the chief medical scientist for IBM’s research arm. In one potential scenario, a patient accesses Watson via a mobile device to explain symptoms in natural language. The person then gets several recommendations for what might be happening, listed in order of the computer’s confidence based on its analysis of relevant data. The patient and doctor would both have access to Watson, which would securely consult medical records and make adjustments to an answer depending on factors such as pregnancy or diabetes, Kohn says.

For now, the features are being developed for care providers, and the effort is still in the testing stage. “There’s going to come a point,” Meyerson says, “where you don’t need any intermediary.”

Before it's here, it's on the Bloomberg Terminal.