The next big UI challenge is making big data human

IBM’s Jeopardy-playing supercomputer Watson is now getting a gig in the retail banking sector as part of an IBM(s ibm) partnership with Citi. This is in addition to its position as a diagnostic assistant for doctors. But the many careers of Watson aren’t just a fun story for the tech press; they illustrate a very big technological and business opportunity for companies like IBM and Microsoft — the rendering of big data into human scale.

For Citi, Watson will be used by retail bankers and loan officers to help them sell to consumers by taking all the information the bank has on a customer and trying to understand what they might want next. So if someone signs up their kid for a credit card, Watson will help the bank understand what type of financial products a customer might want once they send a child to college. For example, my parents remodeled their kitchen. So a Watson-using Citi banker might have offered them a remodeling loan a few months after I left home.

Translating petabytes into insights and then into action is the real value of big data (something we’ll discussing over two days at our Structure: Data conference later this month). And most times in a business setting, charts or spreadsheets aren’t the right way to spark an action. Watson can interface with people using words or speech as opposed to issuing a CSV file.

So while Watson may not speak out loud in his corporate life (he did on Jeopardy), his written statements are delivered in a user-friendly manner and designed to inspire action. However, the power of suggestion is made via voice, in Microsoft’s (s msft) view. In February Microsoft made an equity investment in 247 Solutions, a company that provides the annoying automated voice systems you get when you call a customer service hotline.

But instead of encountering Clippy on the phone, Microsoft wants to use its speech-recognition assets and the cloud to deliver a friendlier user interface for interacting with a lot of data. So when you get a text message from an airline telling you to call a number to reschedule after it canceled a flight, the person at the other end of the phone when you call already knows who is calling, why and how to help.

Unlike another big name in speech recognition, Siri, which uses voice and natural language processing as a means of letting a consumer input a large amount of data and get a result, the enterprise products that Microsoft and IBM are trying use voice and simple commands to filter down a large amount of output into a digestible piece of knowledge.

Manoj Saxena, the GM of Watson Solutions for IBM, explains that the output of Watson crunching through data is designed to be delivered on a mobile phone, tablet or some kind of device that employees carry with them.

“We think the knowledge workers of tomorrow will interface with mobile screens and tablets like the iPad, as opposed to desk-centric PCs,” Saxena said. “So we considered that in building the UI.” He added that later iterations of the partnerships between Watson and various industries will close the loop with the consumer — a person even less likely to have been trained to understand the stereotypical results of a data-crunching supercomputer.

Saxena says they are envisioning an app that will help a patient follow up with a treatment plan for their disease, or help them through financial issues once they get home from the bank or that are perhaps tied to their spending data. So instead of Watson merely schooling Ken Jennings at Jeopardy, he may soon school consumers about their debt-repayment plans or making sure they schedule followup exams. The next big UI question is how Watson will do this without sounding like a nag.

For more on Big Data, including how folks are trying to make it matter to their business, come visit our Structure Data event in New York on March 21 and 22.