Burn Your Best Quarter
Most of what companies call innovation is just velocity. AI made velocity nearly free. What it cannot hand you is the willingness to move against your own business while it still pays.
Motion is not movement
Marketing has never been this busy, or this stuck. Ask a roomful of senior people what they can do now that a competitor can't, and the answers come back as activity. More pilots, faster output, another tool. Almost nothing about where they actually stand.
Most of what we file under innovation does not deserve the word. A new product positioning. A copilot bolted onto the existing workflow. A faster, cleaner version of the campaign you already run. The work looks different and feels like progress, and your position has not moved an inch. You are doing the same job the same way with better equipment, and so is everyone across the street. That is motion. It photographs well and changes nothing about where you stand.
What earns the word does something harder. It is difficult for others to do, difficult to copy once you have done it, and it moves where the value actually sits rather than how fast you cycle through the work you already had. Most companies have very little of it and have talked themselves out of noticing, because the motion never stops and motion is easy to mistake for progress.
None of this is about execution. Execution is doing the work well. Movement is ending up somewhere your competitors can't follow. Plenty of companies are excellent at the first and never reach the second.
The velocity is real. It just isn't movement. AI is a motion machine. It makes the running-in-place faster, cheaper, and more convincing than ever, and almost none of it moves anyone's position, because the engine doing the work is the same one the competitor rents by the hour. Generating options used to cost real money and time. So did producing the work: the drafts, the variants, the test cells, the reporting. This year the price of all of it fell close to zero. The price of turning any of it into a position you can hold did not move.
That is what the disappointment numbers actually show, once you stop reading them as a verdict on the technology. The figure everyone cites, roughly 95% of enterprise AI work producing no measurable financial impact, is directional rather than exact, and it does not describe pilots that failed to run. It describes pilots that ran fine and then stalled on integration, data, governance, and an organization that would not change how it works. The motion succeeded. The position never moved, because moving it was never the model's job.
Where the work piles up
Gary Pisano mapped this years ago. He sorted innovation by two questions: does it need a new business model, and does it need new technical capability. AI made the second one cheap for almost everyone. Capability is rented by the hour now, so the work floods the square that needs neither, the cheap corner where nothing you make stays yours. Every exit from that corner runs up the other axis, through changing how the business itself works, and all of them are expensive. The two squares that need new capability still produce real breakthroughs, but those cluster in a few capital-heavy sectors like chips, aerospace, and pharma, which is why they are not the move most companies can make.
Disruptive
New model, same capability.
Architectural
New model, new capability.
Routine
Same model, same capability.
Radical
Same model, new capability.
- 1BeforeFour squares, two axes. A position could be defended on either.
- 2After AIThe Routine square floods with cheap output: drafts, summaries, reports, forecasts. The work speeds up, the position holds.
- 3The exitStraight up the business-model axis: change who pays, shift what you sell, reach a new buyer. It costs pricing, incentives, decision rights, and capital, and almost nobody makes the climb.
Why anyone still wins
Cheap capability is not only an incumbent's problem, though the incumbent feels it first. The wall that used to protect the big company, years of accumulated capability, a production department, specialist headcount, was the same wall that kept challengers out. It fell in both directions. The collapse that stripped the incumbent's moat also handed a three-person team the ability to do what used to take three hundred. So the honest question is not why the incumbents are exposed. It is why any of them still win.
They do not win on assets. Everyone has assets now, and everyone can rent the same models to work them. They win on something you cannot buy, rent, or download, the willingness to move against your own business while it is still paying. The scarce capability in this whole story is not capability. It is the appetite to burn something that works before a competitor burns it for you.
The willingness to burn
Time Inc. is usually told as a company that failed to see the future. That is not what happened. We saw it clearly. The day it landed, shoppers were standing in the checkout line reading our own articles on their phones while the glossy, profitable magazines beside them sat unbought. We understood exactly what that meant. What we could not do was set fire to the magazine that was still paying the salaries. The assets were all there, the audience, the brands, the data, the distribution. What was missing was the stomach to kill the thing that still worked. The company was taken apart and sold for pieces instead. There were other causes, the debt and the collapse of print advertising among them. The one I keep coming back to is simpler. The failure was not blindness. It was attachment.
The failure was not blindness. It was attachment.
The legacy agencies are living the same refusal now, in slower motion. Their billable core is the production layer, the pure-motion square, and the tools their own clients can run have made it nearly free. Everyone inside those buildings can see it. Almost none of them will cannibalize the billable hour to build the thing that replaces it, because the billable hour still pays this quarter. The data is there. The talent is there. The relationships are there. The resolve is not, and the market is repricing them for it in real time.
Walmart is what the other side looks like, and it is worth watching precisely because it had the most to protect. It is attacking its own retail margin on purpose, turning stores, purchase data, connected TVs, and an AI shopping assistant into an advertising and demand business that competes with the very listings it used to sell.1 Everyone in retail has purchase data. What Walmart has that the others don't is the will to cannibalize its own checkout lane before Amazon does it from the outside. That, not the data, is why it sits on this side of the ledger. Time Inc. owned a comparable hoard and would not light the match.
So the moat was never the assets, and it was never even the ability to move. Movement is the output. The thing that produces it, and the thing no competitor can rent, is the willingness to move against your own book. Attachment is why the assets rot in place. The company that wins is the one that will burn its best quarter to own the next one, which is the exact decision every incentive inside a healthy business is built to prevent.
Which turns the AI question inside out. Everyone is asking what the technology can do. The sharper question has nothing to do with the technology. What is working so well for you right now that you refuse to burn it, even though a competitor eventually will?
Credit. The two-axis map is Gary Pisano's, from You Need an Innovation Strategy (Harvard Business Review, 2015). The reading of what AI does to it, the motion-and-movement distinction, and the argument that willingness is the scarce capability, are mine. The 95% is from MIT's NANDA work on disclosed enterprise deployments, a directional figure rather than a fixed constant.