Field Note · No. 07 · innovation strategy · 9 min read
For most of history, building was the bottleneck. AI moved it — and moving it changes who captures the value, and how you should buy it.
abstract
For a century, the scarce input in innovation was execution. Ideas were cheap and building was expensive, so leaders paid for thinking — decks, roadmaps, workshops — and the value was captured downstream by whoever could actually ship. That arithmetic is what made “innovation” feel like a cost center that occasionally got lucky.
AI changed the arithmetic. Building is no longer the constraint; judgment is. When execution gets cheap and fast, you can stop buying opinions and start commissioning outcomes — and hold innovation on the balance sheet like any other asset, measured and owned. This note makes that argument plainly, names the trap most companies are walking into, and lays out what it means for how an operator should actually buy innovation now.
It is an argument, not a forecast. The one hard number in it — that most corporate AI never reaches the P&L — is sourced. Everything else is reasoning you can check against your own experience.
01 · the bottleneck moved
Every era of business has a scarce input, and strategy is mostly the art of owning it. For most of the last century, the scarce input in innovation was execution. A good idea was cheap — everyone in the room had three — but turning it into a working product took capital, specialists, and time measured in quarters or years. So the market did the rational thing: it priced thinking cheaply and building dearly. Consultancies sold the thinking. Startups and engineering shops did the building, and quietly kept the upside.
This is why so many companies have a drawer full of admired strategy decks and very little to show for them. It was not a failure of intelligence. It was the arithmetic. When building is the expensive part, the value accrues to the builder, and the person who paid for the deck watches someone else capture the market the deck described.
AI moved the bottleneck. The work that used to require a team and a year — the software, the interface, the model, the first version of the brand — can now be produced by one capable operator in a fraction of the time. Building is no longer the scarce thing. Judgment is. And when the scarce input changes, everything about how you should buy innovation changes with it.
figure 1 · the arithmetic flipped
Then
Ideas cheap
Building expensive & slow
Value captured by whoever could build
Now
Building cheap & fast
Judgment scarce
Value captured by whoever commissions the outcome
The scarce input flipped from execution to judgment. Strategy is owning the scarce input — so the question is no longer “who can build this?” but “who is accountable for whether it works?”
02 · the trap: motion without profit
The obvious response to cheaper building is to build more, and that is exactly what most companies are doing. It is also why most of them are getting nothing back. MIT’s 2025 study of enterprise AI found that about 95 percent of generative-AI pilots deliver little or no measurable impact on the P&L; only around 5 percent produce real revenue acceleration.1 The failure is not the models. The report is blunt that the models are fine. The failure is the gap between the thinking and the working system — the same gap that has always swallowed corporate innovation, now decorated with a chatbot.
Cheap building does not fix that gap on its own; it just lets you cross it faster in either direction. Pointed at a real edge with someone accountable for the number, cheap building compounds. Pointed at “we should do something with AI,” it produces a pile of pilots, a line item, and a board asking what happened to the money. The technology changed. The discipline required to convert it into profit did not.
Cheap execution is not an advantage. It is a new baseline everyone will have. The advantage is knowing which thing to execute, and answering for the result.
03 · from buying opinions to commissioning outcomes
If building is no longer the expensive part, then paying for thinking alone — a strategy with no accountable path into production — is paying a premium for the cheap half of the work. The rational move is to stop buying opinions and start commissioning outcomes: define the number the work should move, build the system that moves it, and put accountability for that number on the same person who set it. I call this Active Innovation, and it is only possible now because the execution that used to make it a fantasy has become affordable.
Practically, that collapses the three roles a normal engagement splits across three firms — the strategist who sets the thesis, the builder who ships it, and the operator who has to make the numbers work — back into one accountable arc. Not because one person is smarter than a firm of three hundred, but because the handoffs between those roles are where innovation has always leaked, and cheap building finally makes it possible to remove them rather than manage them. The deliverable stops being a recommendation and becomes a working system, measured against a number set in advance.
04 · innovation becomes a balance-sheet asset
Here is the part most leaders have not priced in yet. When you can commission a specific outcome and measure it, innovation stops behaving like a hopeful expense and starts behaving like an asset you can put on the balance sheet, evaluate, and own. You allocate to it under uncertainty the way you allocate any capital: a portfolio of bets, funded on gates, scaled when they prove out and retired when they do not.
And the same shift creates a second return most owners never collect. Solve a real operating problem and the business is worth more on its own multiple — the first harvest. But if the problem you solved is one your whole industry shares, the fix can become a company of its own, with you holding founding equity in it — the second harvest, on the same work, at no extra cost. One problem, paid twice. That is the thesis behind the Modven Foundry, and it is only rational to pursue once building is cheap enough that spinning the fix into a product is no longer a moonshot.
05 · what it means for the operator
If you run a mid-market company, the practical implications are concrete. First, buy outcomes, not activity: insist that any innovation spend name the number it should move and the person accountable for it, and treat pilots that cannot do both as the cost they are. Second, expect speed — when building is cheap, enterprise-ready work arrives in weeks, not months, and a vendor still quoting quarters is pricing the old bottleneck. Third, run it as a portfolio, not a hero project: govern several bets, keep the optionality to scale winners and kill the rest, and measure everything against a baseline you set before the work starts. The companies that internalize this will treat their innovation function the way good operators already treat capital allocation. The ones that do not will keep funding motion and calling it progress.
06 · the honest limits
Three cautions, because the argument deserves them. Cheaper building does not make judgment cheaper — it makes it more valuable and more exposed, because a bad thesis now ships faster and fails more publicly. This is a case for more rigor at the front, not less. The 95 percent figure is one study of a fast-moving field; treat it as directional evidence of a real gap, not gospel, and measure your own results against your own baseline rather than a headline. And “commission the outcome” is a discipline, not a slogan: it only works if the number is set honestly in advance and someone actually answers for it. Remove that accountability and you are back to buying decks, just faster ones.
references
Put it to work
If you want to point cheap building at a real edge — with a number set in advance and one person accountable for it — the first step is the same as every Modven engagement: one conversation and a written brief, $500, credited in full if we build.