Jeff DeGraff has been called the 'Dean of Innovation' for nearly four decades, a nickname born, he says, as a joke that stuck. DeGraff is a Clinical Professor of Management and Organizations at the University of Michigan's Ross School of Business, and in a new Fortune commentary he makes a blunt argument: most companies don't have an innovation problem. They have a vocabulary problem.
DeGraff traces his own credibility to Domino's Pizza, where he served as Vice President of New Ventures during the company's explosive 1980s growth. He was 27 when the 'Dean of Innovation' title emerged at a retreat with business author Tom Peters. Since then, he writes, he has worked with much of the Fortune 500, the U.S. military, and numerous cultural organizations.
His diagnosis of today's executive suites is specific. In a meeting about artificial intelligence, DeGraff writes, the CEO wants AI as a growth engine, the CIO wants copilots, the COO wants automation, the product chief wants AI-enabled offerings, and the CHRO worries about the workforce. 'Everyone agrees that AI is the future. Everyone agrees that the company needs to innovate. And everyone leaves the meeting talking about something different,' he writes. The result: resources scatter, expectations conflict, and initiatives get measured by the wrong metrics.
Rather than debate whether innovation is 'incremental' or 'radical,' DeGraff proposes a working definition: innovation is 'positive deviance from the norm that creates value.' The two questions that matter, he argues, are what norm is being broken and what new value is created, and for whom. He points to Nvidia, whose graphics processing capability was repeatedly extended into new arenas — most consequentially AI computing — as the kind of departure from the norm that actually creates value, as opposed to innovation used as a corporate aspiration.
DeGraff also names what he calls 'the data trap.' Executives are trained to demand evidence before committing resources, which is generally good management — until it isn't. The more genuinely novel an idea is, he writes, the less reliable historical data becomes. Demand too much proof and an organization unintentionally selects for the familiar; innovation quietly becomes optimization. He cites generative AI: in late 2022, no executive could produce a credible five-year ROI model because the technology, costs, competitors, regulations and use cases were changing too fast. 'The companies that waited for certainty did not reduce uncertainty. They simply learned more slowly,' he writes. His alternative is to ask not whether an idea can be proven, but what is the cheapest, fastest experiment that reveals something important.
The pattern DeGraff describes inside boardrooms should sound familiar to anyone who has watched a regulatory agency demand certainty before permitting a market to move. Committees that require proof before capital gets deployed do not eliminate risk — they simply relocate it, favoring the safe and the familiar over the genuinely valuable. The market, unlike the internal committee, does not wait for a five-year model. It prices the experiment, rewards the deviation that creates value, and lets the rest go. Capital, in the end, rewards clear rules and real bets — not consensus about a buzzword.
