Before the capital moves, an independent read of whether the AI is real.
For partners, directors, and owners committing capital against an AI claim. The deck is convincing. It was built to be. The question is what is underneath it.
The question
A portfolio company reports its AI is going well. A target calls itself AI-first, years ahead of the market. Sometimes that is true. Often what exists is a set of pilots arranged to look like a strategy.
The difference is rarely visible from a management presentation. It sits in the data foundation, in who actually owns the AI decision, and in whether anyone in the company has the authority to change what the business aims at. I read those directly. I spent ten years building the layer these claims stand on, and I know what it looks like when the ground will not carry the plan.
Some of what I flag runs against instinct. "We have built this technology over ten years" sounds like depth. Often it is a cost: a stack that age usually needs fresh investment before it can carry what the pitch promises, and nobody can tell you how much until someone reads it. Better before the commitment than after.
How we can work
Scoped to the decision in front of you, not assembled from a package.
-
Independent Diligence Read
Before the commitment: the AI strategy, the execution claims, vendor dependency, key-person concentration. The elements that do not appear in a management presentation.
-
Portfolio AI Strategy Audit
For an existing investment: is the AI connected to business outcomes, is execution on track, is the capital going where it should. And if the goal is to raise the company's value through AI, what that would actually take.
-
Startup and Deal Evaluation
An independent technical voice when an AI-driven company is in front of you: whether the claims hold, and whether the team can do what the pitch describes.
How I work
First, the situation as it is: what is claimed, what actually runs in production, and what decision sits in front of you. Then validation: does the strategy connect to outcomes, do the claims survive contact with the work. Then risk: vendor dependency, key-person concentration, strategic fragility. Then a written recommendation: keep, change, or stop, with the reasoning attached. My job is to inform a decision, not to justify one already made.
You walk away with a document you can put in front of an investment committee or a board, not a slide deck.
Diligence rarely waits. I work to the deal clock, and a short scoping call settles what is possible inside your window.
Where I stand
I do not come from the deal side. I come from the side that built the systems the deals are priced on. The read is independent by structure: no vendor to recommend, no implementation contract waiting at the end.
How I think about AI in mid-market companies, the companies portfolios are made of, is published and open: Velocity with Containment: An AI Adoption Doctrine for Mid-Market Leaders
Others put me in the judging seat: I judge AI startup pitch competitions, most recently the Everything AI competition at New York Tech Week. Speaking and Judging
If capital is about to move against an AI claim, describe the decision and the window. A thirty-minute call is enough to scope the read, the timeline, and the cost.
Start here