Summary:

IBM’s always on the look out for new challenges for Watson to tackle. Two dozen teams of USC students recently had 48 hours to create their own business plans for the technology.

Watson
photo: IBM

IBM’s Watson has already proven it can beat human brainiacs at Jeopardy and has shown promise in cancer research. It has even gone to college at Rensselaer Polytechnic Institute. Now, results of a new academic challenge at the University of Southern California may show that Watson can also suck cost out of legal proceedings and help sufferers of post-traumatic stress disorder (PTSD).

Two dozen student teams participating in the recent IBM Watson Academic Case Competition at USC had 48 hours to come up with a new use case along with a business plan that would harness Watson’s natural language processing skills and cognitive capabilities. The plans were judged by a panel of IBM execs, school officials and business leaders.

“The goal is to come up with a business plan that had to be feasible, well thought-through, come up with a go-to market, and it had to have a stable business model,” said said Steve Gold, VP of Watson Solutions for IBM. IBM provided a crash course on Watson’s capabilities, one-on-one consulting and an open Q&A — and then turned the teams loose.

IBM wanted students to work across schools and areas of expertise. Its stated goals were to expose new people to Watson’s capabilities, advance the curriculum around Watson and encourage research that could help build out Watson’s talent pool.

Putting Watson to work in law, training and PTSD research

The first-place team came up with a plan to use Watson in a legal setting, an area some would say is rife for disruption. Law firms traditionally relied on per-hour fee arrangements and those fees quickly add up. Now, law firms are “encouraged” to use a flat fee arrangement, Gold said.  The outsourcing of legal work — when a law firm ships work off to a third-party provider but bills clients for that service,  is already a $4 billion business, Gold said. All those legal documents are a perfect example of the reams of unstructured data that Watson is great at parsing. In January, GigaOM’s Derrick Harris touched on the legal discovery as a  big data application. 

This team recommended that Watson be used to conduct discovery for corporate legal departments — sifting through court documents, briefs, legal articles and related material.  The takeaway, according to IBM:

“By placing Watson in charge of research, firms can recover time and costs while delivering better legal outcomes. In turn, firms that leverage Watson’s speed and efficiency can address the legal trend towards ‘flat fee’ billing and research outsourcing.”

The second-place team recommended that corporate human resources departments use Watson to evaluate data about employee career goals and assess and recommend the training options needed to attain them. In theory, a successful application would mean better-prepared employees and a higher level of job satisfaction.

The third-place team proposed that doctors use Watson to find undiagnosed PTSD patients by sifting through military veterans’ data, including their medical histories and combat records. The problem with PTSD is that it often is not reported by the sufferer.

None of the USC teams actually had access to Watson, although that could happen down the road, Gold said. “There’s only so much you can do in 48 hours. The goal here is to find a market that needs addressing, and deliver a plan to attack it that would be applicable to Watson’s talents,” he said.

Next year, IBM and USC hope to expand the competition to 500 students across USC’s various schools. IBM has run similar Watson competitions at Cornell and the University of Rochester. One thing’s for sure: IBM will milk Watson’s success for all it’s worth and also use it as a recruiting tool. Last year, Gold’s group brought on 17 interns and plans to field 26 to 28 interns next year, including two high school students.

You’re subscribed! If you like, you can update your settings

Comments have been disabled for this post