Ginni Rometty on research, social, and Watson

October 16, 2013, 3:54 PM UTC

FORTUNE — Ginni Rometty’s advice to companies: Don’t define your business by a product. In a world of consistent and constant change, it’s the most dangerous thing for a company to do. “You’ll miss [technology] shifts. You’ll miss dangerous ones like business models,” said Rometty.

As Chairman and CEO of IBM (IBM), Rometty has seen the danger firsthand. Having spent more than 30 years at the company, she lived through the near-death experience of the early 1990s and is now just the ninth CEO in the company’s 102-year history. It’s this role that has earned Rometty, 56, the top spot on Fortune‘s list of the Most Powerful Women in Business.

MORE: Complete coverage of the Most Powerful Women Summit

Speaking October 16 at the Fortune Most Powerful Women Summit in Washington, D.C., Rometty outlined five areas on which she depends to help her prepare IBM for future markets. First, there’s research, an area in which the company has long been a leader. The company employs 3,000 Ph.D.s — the largest research outfit in the technology industry — and has been the global patent leader for the past 20 years. Just as important are relationships with clients, who are increasingly partnering more closely with IBM on innovation initiatives. Rometty also nurtures relationships with universities and with the venture capital community. And, thanks to new social tools, Rometty has been able to connect more deeply with IBM’s 400,000 employees to unearth ideas bubbling up internally.

One secret to IBM’s success has been the discipline it applies to nurturing these relationships and harvesting insights from them. For example, Rometty described the company’s global technology outlook, a document that researchers prepare annually to map out future trends. Rometty depends on it to inform strategy, and each year grades past GTOs for their success in predicting trends.

It’s this discipline that leads to breakthrough computing paradigms like Watson. It’s been two-and-a-half years since the Jeopardy-winning computer beat a couple of brainiacs on national television — and demonstrated that computing tools could store and sift through copious amounts of information, respond to natural language, and learn from its mistakes. It’s technology that represents decades of work in IBM’s labs, and the company is just beginning to commercialize it. I wrote about IBM’s efforts to bring it to market in a recent Fortune feature .

Now, says Rometty, “Watson is off to medical school.” The technology will be deployed at the Cleveland Clinic to help medical students better understand medical literature.

Meanwhile, in IBM’s labs, scientists are working to improve Watson. “A Watson 2.0 will see — pictures of things,” says Rometty. “And a Watson 3.0 can reason with humans.”