When AI advisory becomes co-invest.

Most advisory engagements bill in dollars and end when the deliverable lands. A small fraction don't — they produce something the client could plausibly sell to other companies, and the question of who owns that capability becomes the most important question in the engagement. GenovateAI's co-invest structure is the answer we use when this happens. It is not a default. It is offered, never required, and most engagements never trigger it.

This essay exists because the co-invest line on the homepage is the most-asked-about and least-understood part of the offer. CFOs see it and immediately want to know what it means for fees, equity, and exit. CSOs see it and immediately want to know what it means for IP, employment, and competitive risk. Both sets of questions are reasonable. The standard advisory engagement is the right answer in most situations; the co-invest engagement is the right answer in a few specific ones, and the rest of this piece is about how to tell the difference.

What it actually is.

The co-invest structure is a modification to the Build and Embed phases of a standard engagement. Up to and including the Design phase, the engagement runs identically to any other — fixed fee, fixed deliverables, no equity component, no obligation on either side.

The co-invest conversation happens at the end of Design, when both sides have a clear picture of what is going to be built. If the capability has obvious value beyond the engaging client — that is, if other companies in similar situations would plausibly buy or license it — the engagement can convert into one of two structures. The first is a partial fee-for-equity swap: a portion of Build and Embed fees is exchanged for an equity stake in a new entity that owns the productisable capability, typically in the 0.5%–2% range, with the engaging client as the founding customer. The second is a revenue share: fees remain fixed but a small share of any commercial revenue from third-party deployments of the capability flows back, typically 3%–10% for a defined period.

Either structure assumes a spin-out — a separate legal entity owns the productisable capability, with the engaging client as anchor customer and a defined IP arrangement that preserves their use case forever. The engaging client never loses access to what was built for them. They share, by election, in the upside if it turns out to have broader value.

The four conditions, all of which must be true.

Co-invest is the wrong structure for most engagements. The four conditions below all need to be present for it to make sense for either side. If any one is missing, the standard advisory engagement is the right answer.

The matrix is deliberately strict on the "wrong" side. In practice, roughly one in five engagements meet all four conditions for co-invest. The other four-fifths run as standard advisory. The point of having the structure available is not that it gets used often — it's that when it should be used, both sides have a clean framework for it instead of inventing one mid-engagement.

The CFO's real question.

The question CFOs actually ask, once the structure is clear, is simpler: "Does this change how the advisor behaves during the engagement?" The honest answer is yes — and the way it changes is the reason the structure exists.

In a standard advisory engagement, the incentive structure points the advisor toward delivering the scoped work on time and on budget. There is no built-in reason for the advisor to push for the harder, longer, more architecturally clean version of the work — that version costs more time and produces the same fee. Good advisors push for it anyway because it's the right work; the structure does not help them do so.

In a co-invest engagement, the incentive structure points both sides toward the longer-term, harder, more durable version of the work. The advisor's upside lives in the capability holding up under commercial scrutiny from other potential customers; this is also exactly what the engaging client wants for their own use. The interests align, not at the level of pleasantries, but at the level of architectural decisions made in week eight.

The CSO's real question is different, and equally legitimate: "What stops the advisor from taking what they learned about my company and selling it to my competitors?" The answer is contractual, not philosophical. The spin-out entity has explicit non-compete and confidentiality terms with respect to the engaging client's domain and customer set. Anonymised pattern learning is fair game; specific competitive information is not. These terms are written into the spin-out shareholder agreement, signed by both sides, and have teeth.

Why we offer it at all.

Two reasons. The first is that the structure produces better engagements when it fits — for the reasons above, with aligned incentives at the technical layer where the work actually happens. The second is more candid: GenovateAI sits inside Meta3Ventures, which is a venture studio and fund. The infrastructure to identify, structure, and operate spin-outs already exists at Meta3. Offering it to engagement clients is using infrastructure we built for other reasons.

The structure is not for everyone. It is not appropriate for most engagements. It is offered when conditions are met, declined cleanly when they're not, and never used as pressure to choose us over another advisor. If a co-invest conversation feels in any way pushed, that's a signal it's the wrong engagement — for both sides — and the right call is to either run it as standard advisory or to walk away.

The most important sentence in this entire piece is the simplest one: the co-invest mechanism is optional. Most engagements never trigger it, run beautifully without it, and end on standard advisory terms. The reason it appears on the homepage is that it's an honest description of how we work when it fits — not an aspirational tagline meant to differentiate. If your engagement doesn't fit, we'll tell you, and you'll get the standard structure with no part of the offer changed.