The playbook was always there. AI just raised the stakes.
Alpha must be made now, not bought. What the consensus leaves unsaid is that it is made or missed on phase. The value thesis was always a lifecycle — Transform, Grow, Hold, Exit — and each leg carries a different bet: a different tolerance for risk, a different return on the same capability. AI did not add a lever. It raised the price of using the levers you already have out of phase. Twelve where five used to do is not a technology problem. It is a phase problem.
Every value-creation plan is underwritten against something. The default is to underwrite against the technology's cycle — the capability that is ascendant, the one the board walked in asking about — instead of the asset's lifecycle. That is the error. The bet has to be underwritten against the asset's phase, and the return that phase was bought to produce. Drop a transform-era bet into a hold and it does not just miss the base case. It works against the equity story, spending the de-risked cash flow the thesis underwrote for the exit. Alpha is not a function of how advanced the capability is. It is a function of whether the phase can bank it.
The structural remake. Reset the cost base, redraw the operating model, rebuild capability. The biggest AI bet lives here, and it is the leg most often faked: underwritten as transformation, funded as a migration.
AI as an EBITDA multiplier. Margin, pricing, productivity, cost-to-serve. It has to show up in the numbers, because Grow is where the alpha is banked.
AI as efficiency and de-risking. Incremental, governed, protecting the cash flow the exit is underwritten on.
AI as a diligence narrative. Legible and defensible in the data room, a credit to the multiple, not a red flag.
The capability is the same on every leg. The bet is not. Matching the bet to the phase, leg by leg, is not a tooling decision. It is an architecture decision, and it belongs to a seat most cap tables do not have.
Every failure across the four legs is the same failure. No one owns the phase.
The CIO keeps the stack running. The CTO builds the product. The operating partner owns the number. The Chief Transformation Officer owns the program — the plan, the milestones, the number it has to hit. The board underwrote the thesis. Every one of those seats is real, and filled, and not one of them owns the translation of the thesis into a sequenced, phase-aware target state, or runs the remake the Transform leg was supposed to be.
The obvious objection is the transformation officer. That seat owns the program, not the architecture. Owning a plan is not owning its translation into a buildable, phase-sequenced target state, and it is not running the remake. The officer cannot build it. The CTO does not own the arc. The work that joins them is improvised on every deal, and the improvisation is why Transform gets funded as a migration.
That gap is its own seat. The Chief Transformation Architect.
It owns what neither side does: matching the bet to the phase, translating the thesis into a target state, and sequencing the remake so each leg sets up the next. Not the plan, the architecture. Not the migration, the remake. It is the seat that makes twelve reachable instead of underwritten and missed.
Most cap tables do not have it. That is why the bets fail.
Advisory to private equity firms and the companies they hold. I take a small number of engagements at a time. The first step is a conversation.
me@andrewphunter.com →