Ask a frontier model what your AI strategy should be. Give it your industry, your revenue band, your competitive position. What comes back in 2 minutes will be coherent, specific, and sequenced about as, or at least nearly as, well as a competent consultant would have produced for forty thousand dollars two years ago. It will name the right use cases. It will flag the obvious risks. It will even tell you which two projects to start with.
That should, in theory, terrify anyone who sells AI strategy for a living. For Innovation Vista, it really doesn’t; and the reason it doesn’t is the entire subject of this piece.
The answer to “what should we do about AI” has collapsed in price. Not declined; collapsed. The thing mid-market CEOs spent the last three years anxiously shopping for, the strategy, the roadmap, the prioritized use-case list, is now something a model will hand you for free and a competitor will hand you in their next press release. If the answer were the product, this would be an extinction event for the firms that sell it.
It isn’t, because the answer was never the expensive part. It was never even the hard part. The expensive part has always been adoption: the slow, political, irreducibly human work of getting an organization to actually do the thing the answer describes. That has always been true. It is more true now than it has ever been, and the gap between the two is widening every quarter.
The Answer Was Always the Cheap Part
There is a version of the last thirty years where a CEO could be forgiven for believing the answer was the scarce thing. For most of the enterprise-software era, the people who held the answers also held the gate. You couldn’t get the roadmap without buying the engagement; you couldn’t get the engagement without the budget cycle; you couldn’t run the budget cycle without the deck. The answer felt expensive because access to it was rationed.
So mid-market leaders learned to behave as though the answer were the prize. They benchmarked vendors on the quality of the recommendation. They graded consultants on the polish of the strategy document. They treated the arrival of a good plan as the moment the problem was solved, when it was actually the moment the real problem began.
That posture was defensible. It was even rational, given how the market was structured. But it quietly trained a generation of executives to over-value the cheap input and under-value the expensive one. The deck got the standing ovation; the eighteen months of organizational grind that determined whether any of it mattered got a line item and a sigh.
Anyone who has actually run a transformation knows the truth in their gut: the plan was never what kept you up at night. What kept you up was whether the regional VPs would use the new system or quietly keep their spreadsheets; whether the incentive structure rewarded the behavior the strategy assumed; whether the thing would still be running the way you designed it eighteen months after the consultants left. The answer was the easy forty hours. Everything after the answer was the hard two years.
What Changed Is the Price of the Initial Answer, Not the Price of the Result
For three years, waiting was a coherent strategy. The models weren’t reliable enough, the tooling wasn’t mature enough, and the smart move was to let the early adopters absorb the cost of the bleeding edge. We have made that argument ourselves, and we still think it described the period honestly: there were real reasons that waiting felt like prudence rather than paralysis.
What broke that posture is not that the answers got better. It is that they got nearly free. The capability that rebuilt a frontier AI company is now available off the shelf at roughly the price of electricity; the recommendation engine that used to live inside a consulting firm now lives inside a thirty-dollar subscription. The marginal cost of a competent strategic answer is approaching zero, and it is not coming back up.
Here is the part most executives miss, because it runs against the instinct the last era trained into them. The price of the answer falling to zero does nothing to the price of the result. A free, perfect strategy document still has to be absorbed by an organization full of people, habits, incentives, and legacy systems that did not get any cheaper to change. If anything, the cheaper the answer gets, the more conspicuous it becomes that the answer was never where the money was going.
This is the same diagnosis we have made about why AI initiatives stall. The projects are not failing because the models are insufficient; they are failing because the operating model around the models never changed. The answer was sitting right there, correct and unused. RAND puts the AI project failure rate north of eighty percent, roughly twice the rate of ordinary IT projects, and almost none of those failures trace back to a bad answer. They trace back to an organization that received a good answer and could not metabolize it.
Where the Cost Actually Lives
It helps to be precise about what “adoption” means, because the word gets used as a synonym for “rollout,” and rollout is the cheap, scheduleable part. Adoption is the part that resists scheduling.
Adoption is the regional manager rebuilding the comp plan so the new workflow is the one that pays. It is the three legacy systems that each call the same customer something different, so the AI has nothing clean to act on until someone does the unglamorous reconciliation work. It is the governance question of who is allowed to let an agent touch production, and what happens the first time it is wrong. It is the forty employees who each quietly built their own version of the tool, so that what you have is not one capability but a rack of mismatched, ungoverned reinventions running critical workflows nobody can see. It is the eighteen-month stretch where the strategy is technically “in flight” and practically indistinguishable from nothing having happened at all.
None of that can be bought off a shelf, because none of it is an information problem. It is an absorption problem. You cannot prompt your way out of a misaligned incentive structure. The model cannot reconcile your master data for you, not because it lacks the intelligence but because it lacks the authority, the institutional knowledge, and the standing to make forty people change what they do on Monday morning. Adoption is expensive for exactly the reasons it has always been expensive: it is human, it is slow, it is political, and it does not compress no matter how fast the underlying technology moves.
This is why the firms whose entire value was the answer are quietly the ones most exposed right now, and the firms whose value was always the absorption are the ones with the most durable position. The market is about to find out which is which.
Why Adoption Got More Expensive Exactly As Answers Got Cheaper
There is a counterintuitive turn here worth pointing out. You would expect that as answers commoditize, the whole problem gets easier and cheaper…. but the opposite is happening. The cheaper the answer, the more of the total cost concentrates in adoption, and the more decisive adoption becomes as the thing that separates winners from losers.
The logic is simple once you see it. When the answer was scarce, having the better answer was a real edge; the company with the sharper strategy could win on strategy alone – EVEN if they suffered some implementation challenges. Now everyone in your sector can get at least get to a “competent” answer in an afternoon. The strategy document is no longer a differentiator, because your three closest competitors are holding a nearly identical one. The only thing left to compete on is which of you actually absorbs it, and how fast.
That is the same arithmetic underneath our argument that within a decade you will be able to buy competitors for pennies. The repricing that is coming does not accrue to whoever knew what to do; everyone will know what to do. It accrues to whoever did it. The gap between the company that has the answer and the company that has banked the result is no longer a gap in knowledge. It is a gap in adoption, and it compounds. The leader is not eighteen months smarter than the laggard; the leader is eighteen months further into the grind that can’t be skipped, and the laggard cannot buy that time back at any price, because the cost was never denominated in money.
So the proposition in the title is not a clever inversion. It is a description of where the cost has moved. The answer is cheap, and it always was; what is new is that it has gone from cheap to free, which means one hundred percent of the cost, and one hundred percent of the advantage, now lives in the part of the work that no one can buy their way past.
What It Looks Like to Take Adoption Seriously
If the answer is free and the absorption is everything, then the only work worth paying for is the work that happens at the adoption layer; and that work looks almost nothing like a strategy engagement.
It looks like an outside leader who is in your business on a weekly cadence, not a quarterly one, because absorption happens between the meetings and not in them. It looks like someone who arrives already fluent in your sector’s terminology, regulatory landscape, and vendor dynamics, so that the first month is spent reconciling your data rather than learning what your data means. It looks like the willingness to stay through the eighteen-month stretch where nothing visible is happening and the only thing being built is the boring, durable plumbing that determines whether any of it sticks. It looks like governance designed as a freeway rather than a roadblock, so that the absorption accelerates instead of stalling at the first hard question about who gets to touch production.
And it looks, most tellingly, like a firm willing to put its own money on the same outcome the client is betting on. When the answer was the product, the incentives could afford to be lazy: the firm got paid on delivery of the deck, and whatever happened during adoption was the client’s problem. When adoption is the product, that arrangement is incoherent. You cannot credibly sell absorption while structuring your fees so that you are paid in full before the absorption begins. The structure has to follow the thesis. If the expensive, decisive part is adoption, then that is the part you have to stand behind, with skin in the game, until the result is banked and not merely recommended.
We are not describing an aspiration. We are describing where we have chosen to stand, because it is the only position consistent with the diagnosis. More than half of the measurable impact across our engagements has come from the top line rather than from cost savings, which is the kind of result that only shows up when someone stays past the answer and into the absorption. The answer is the part we will, frankly, give you in the first conversation. The reason to keep talking after that is the part the first conversation cannot do for you.
The Cheap Part and the Expensive Part
The hardest habit to break is the one the last era trained into so many of us: treating the arrival of a good plan as the finish line. It was always the starting line, and the gun has just gone off in a race where everyone is holding a map of the course.
If you take one idea from this into your next board meeting, let it be this pivot: the board does not need to ask whether you have an AI strategy; you can produce one before the meeting ends, and so can every competitor in the room. The question worth the board’s time is whether you have the capacity and the leadership to absorb it faster than the company across the street. That is the scarce variable. That is the one with a price tag that has not fallen and will not fall.
The answer is cheap. It has never been cheaper. Adoption excellence is what costs you, it has always been most of what costs you, and the only thing different now is that it’s the only cost line item on the invoice. The companies that internalize that this year will spend the next decade buying the ones that didn’t.
