Watson, IBM’s artificial intelligence computing platform, is changing the way we compute. From its roots as a robotic contestant on Jeopardy, the machine-learning marvel is now being positioned as a tool for doctors, businesspeople, and scientists worldwide--one that can answer any question posed to it in natural English.
The latest example is a partnership between IBM and the military financial service USAA. Together they're introducing a Watson-based advisor for veterans returning to civilian life. By visiting USAA.com, veterans can ask Watson finance-related questions about their post-military transition. IBM says that USAA is the first to commercialize Watson into a consumer-based application..
IBM’s future arguably depends on Watson's success. The company’s hardware sales have been falling for the past few years, and this computer system has been their most notable growth area. In January, the IBM invested $1 billion in a massive expansion of Watson’s ecosystem
Mike Rhodin, the senior vice president and effective head of IBM’s Watson group, is the man in charge of making sure that investment yields returns. Fast Company had the opportunity to speak with Rhodin late last week. Our first question was about the recent wave of speculation that IBM and Apple would engage in a series of Watson-Siri collaborations following news of a landmark corporate agreement between the two former (fierce) rivals. For both Apple and IBM, a merger between the two projects would make sense in many ways. He declined to comment, on any Watson-Siri partnership, But did give us a window into Watson's road map.
Law, media, and health care came up multiple times in our conversation as a focus of Watson's intelligence-gathering. IBM sees the platform as an antidote to information overload. Referring to law and education as examples, Rhodin said that “The amount of information created within those industries overwhelms professionals’ ability to consume it. We believe that is the catalyst for innovation.”
To do so, IBM is feeding Watson massive amounts of texts and records to make it “smarter.”
“Doctors are trained on a set of information in med school and go through internships and residencies. But when they’re practicing, they only have so much time to catch up with new information. When you add in something like low cost DNA sequencing in genomics, it's simply overwhelming,” Rhodin says. Doctors and nurses are already using Watson in the field, and soon it will become an even more reliable advisor.
While Siri is aimed for the home and the car, Watson (for now) is aimed at the workplace. IBM’s job listings boards offer hints of where the company intends to take the product; one recent posting was for a meteorologist summer intern with experience in the energy industry. Rhodin confirmed that IBM is considering the oil and gas industries as potential Watson consumers. Another job listing for a product manager confirms that IBM wants to position Watson in government and the financial sector. Rhodin cites wealth management as one category he sees a particular niche for Watson in.
As a company, IBM is far more comfortable dealing with enterprise customers and large corporate or institutional clients than the consumer market, which it traditionally has had issues reaching. Early Watson efforts have been concentrated on health care, which is a perfect example of an industry dominated by relatively few institutional players. Rhodin told me the company wants to hire anyone who, as he puts it, “has domain level expertise.”
Rhodin says IBM’s recently created Watson group was formed by the largest transfer of research personnel in the company’s history. “One of the things we did with the creation of the Watson Group was creating a reasonably autonomous unit in IBM that integrates research, consulting, delivery, and sales in one unit that delivers to me. I have the equivalent of a small company running the business, with the ability to adapt and move quickly,” he added.
The backend of Watson--the servers, software architecture, and API which allow developers to build apps, relies on a process called “cognitive computing.” In layman’s terms, cognitive computing allows software to mimic perceptive, cognitive, and interactive aspects of the human brain.
Earlier this month, IBM announced a $3 billion R&D investment in computer hardware that mimics the human brain. In an increasingly cloud-driven world where diverse arrays of companies rely on a remote infrastructure (See: Amazon Web Services, Salesforce, and the Google ecosystem), IBM is positioning themselves as a major player for cognitive software.
Equally important is the fact that IBM is the only major player in the cognitive computing movement. While several smaller companies exist working in the field, IBM has been the industry leader ever since they got a grant from DARPA (along with five universities) to conduct cognitive computing research back in 2008.
IBM held a Watson mobile challenge at this year’s Mobile World Congress as a way of finding case studies for Watson outside of desktop computers. The three winners were a tablet based trainer for in-store retail personnel called Red Ant, a personal health care wellness assistance tool called GenieMD, and a company called Majestyk Apps which made a prototype stuffed animal called FANG (Friendly Anthromorphic Network Genome).
“FANG is a toy you give to little children that connects to Watson,” Rhodin told Fast Company. This toy talks back, but what’s interesting is that the parents have an app that actually tracks conversations between children and the toys. It seems creepy at first, but when you step back you realize they are important developmental years for the child. Analytics can help parents track development of children. You can look at this either as parents trying to over-engineer their kids, or you can look at it from a much more important viewpoint of developmental progression and detecting how there might be problems.”
There has been speculation by many industry observers that the recent Apple-IBM partnership could ease the way for iOS developers to work in the Watson ecosystem. But as we’re about to see, IBM is already laying plans for a massive Watson ecosystem.
One thing Rhodin seemed especially happy about during the interview was IBM’s work building partnerships with universities to steer developers towards Watson. Rhodin told Fast Company that this fall, 10 U.S. universities would begin offering Watson-based computing classes including Carnegie Mellon, Ohio State, the University of Texas-Austin, the University of Michigan and New York University. As he put it, “A large number of top schools in North America will train people on how to build cognitive applications and be the next generation of cognitive entrepreneurs in market.”
IBM has also been giving outsiders access to Watson’s API (though applicants have complained of a glacial pace in approvals) to build out applications across a variety of industries. “Our biggest issue right now is that so many people came out to us that we had to gate access,” Rhodin added, “And scale up the backend. For those who want to work with us up front, there’s no up-front charge, just go through the qualification process so we can hook you up with the system, show you the API, and have you up and running. It's a revenue-sharing program going forward, if our licensees don’t succeed, we don’t have anything either.”
[Image: Flickr user Idaho National Laboratory]