A.I. isn’t just some shiny object to feed futurist dreams—it’s critical to building a better health system and understanding health across a wide population, argues IBM executive vice president John Kelly.

Kelly spoke Tuesday at Fortune‘s Brainstorm Health conference in San Diego as part of an expert panel on machine learning and cognitive computing’s role in medicine. “I am wildly optimistic but very practical about it,” said Kelly, referring to the use of big data analytics in the field.

But he, like fellow panelists Mona Siddiqui, chief data officer for the U.S. Department of Health and Human Services, and Gianrico Farrugia, president and CEO of the renowned Mayo Clinic, offered some key caveats.

“The quality of the data, the size, et cetera, all of these things matter,” Kelly said, adding that unlocking actual insight from this avalanche of information is so complex that humans simply can’t do it alone. “A.I. is the only way to deal with the mass of data that we’re collecting.”

IBM has a vested interest in doing just that through its Watson supercomputing unit. The tech giant’s Watson Health has faced some critical media reports in recent years on how much it has tangibly accomplished. (Asked where Watson currently stands, Kelly replied: “We’ve made tremendous progress with IBM Watson in health care.”)

That includes helping insurers and other payers better design benefit plans by poring through claims databases, matching cancer patients to clinical trials, and helping doctors come up with treatment plans for their patients, according to Kelly.

“Matching a cancer patient with a clinical trial takes a long time,” he added. “Through Watson and our partnership with the Mayo Clinic, we’ve helped double [certain] clinical trial participation rates.”

For more coverage of Fortune’s Brainstorm Health conference, click here. For news delivered daily to your inbox, subscribe to Fortune’s Brainstorm Health Daily newsletter.