IBM targets financial firms with commercial Watson

IBM has laid out a commercial roadmap for its supercomputing system Watson after inking an “exploratory agreement” with US financial services giant Citigroup this week.

Under the agreement, Citi will use Watson’s analytics and evidence-based learning capabilities to extract information about customers and risk from its unstructured data.

Citi planned to incorporate Watson in a “first-of-a-kind customer interaction solution” that could suggest products or services based on customers’ individual circumstances.

In February last year, Watson famously beat human opponents in the US quiz show ‘Jeopardy!’ using about $3 million of hardware and custom code to cope with the format of the show.

US health insurance company WellPoint became Watson’s first commercial client in August, with IBM agreeing to apply the system to the insurer’s historical data of treatments and outcomes.

IBM Watson general manager Manoj Saxena told the vendor’s Pulse conference this week that it was applying Watson and its 41 subsystems to cancer research.

He said IBM had developed Watson “advisor cartridges” for oncology, diabetes and cardiology to combine deep domain expertise with its technological capabilities.

The vendor planned to build similar cartridges – such as a “financial advisor cartridge” – with Citi.

“Watson will be delivered as a cloud-based service and we will be pricing it as a portion of the value that we deliver to our clients,” Saxena told the conference.

“The cloud deployment could be either a private cloud, public cloud or a hybrid cloud,” he said, noting that public cloud services would be faster to deploy and more scalable, while private cloud services allowed clients to comply with any regulations.

“This is our five-to-seven year road map.”

Saxena said clients had demonstrated a “tremendous amount of interest in Watson” but many of those were either more suited to other analytics solutions or did not have infrastructure in place to support the technology.

He planned to address the latter issue with IBM’s new “Ready for Watson” integration path that would take clients on a “multi-year journey” to build up their big data and content analytics capabilities.

The Watson business would focus on English-speaking healthcare and financial services sectors this year, Saxena said, highlighting future opportunities in contact centres and governments.

“Industries like healthcare, telecoms, banking, insurance – where information liquidity is what drives business outcomes and business effectiveness is what we’re focusing Watson on,” he said.

“Computer systems for the last five decades have been built around structured data and structured systems,” he said, explaining that this excluded unstructured data like human speech.

“Watson is the beginning of new technology that starts working on top of structured, semi-structured and unstructured data.”

In Australia, the Commonwealth Bank has begun working with research organisation NICTA on a Hadoop data mining platform that will allow it to offer loans based on individual customers’ habits.

NICTA researcher Rami Mukhtar expected big data analytics to be shaping Australian retail banking products, particularly over internet banking, within the next 18 months.

IBM Pulse attracted some 8000 attendees, including a number of Australian clients that will tour IBM’s Innovation Center in Dallas, Texas, later this week.

The tour will feature a Watson demonstration, although researchers and the infrastructure on which it operates is based in New York.

Liz Tay travelled to the Pulse conference in Las Vegas as a guest of IBM.