The Assignability Problem

Observability is not assignability. We treated those as one thing. They are not, and the gap between them is the whole argument.

A few weeks ago we argued, in a companion piece, that AI is quietly turning professional services into a financeable asset. The chain was simple. Machine-readable, auditable, underwritable, financeable, securitisable. Make the work legible and the capital follows. The chain had a missing link.

The people who pushed back hardest were the ones worth listening to. Not the consultants, who mostly said it was mind-blowing. The bankers and the structured-credit people, who said it was interesting, and legally messy. They were right. The mess is the point.

Observability Is Not Assignability

We treated "you can see the cash flow" as if it meant "you can claim the cash flow." Those are different things. The gap between them is where structured finance actually lives.

A mortgage securitises because three things are true at once. The cash flows are contractual. The repayment rights are transferable. The collateral is legally enforceable. AI can deliver the first for services. On its own it does nothing for the other two.

Ask the question a credit committee asks. A client hires a firm for an eighteen-month transformation. The firm fails at month nine. Can the lender seize the work in progress? Can it swap in another team and keep the engagement alive? Is the contract even assignable without the client's consent? The answer is usually no, no, and no.

Services cash flow behaves less like a mortgage and more like key-person equity. The asset and the people are the same thing. The people go home every night.

Insert the link we skipped. The real spine is machine-readable, auditable, assignable, underwritable, financeable. Assignability is the load-bearing step. The barrier was never observability. It is the assignability of economic rights independent of the people who deliver the work.

You Cannot Repossess Expertise

This is the deeper version of the same problem. A bank likes an asset with residual value, something to recover when things go wrong. Consulting work in progress has almost none. You cannot foreclose on a partner's relationships. You cannot foreclose on tacit knowledge. You cannot foreclose on the credibility that wins the next mandate. Those walk out the door and do not come back.

A law firm's pipeline is a set of promises held in specific heads. Repossess the office and you have repossessed furniture. This is the part that makes "asset class" harder than we wrote it. An asset you cannot recover is an asset a lender prices with fear.

Finance Has Done Exactly This Before

And yet it has. Finance has a long history of learning to monetise cash flows that looked unpriceable and unrecoverable. It managed it every time by isolating the claim from the mess underneath it.

Invoice factoring monetises receivables despite execution risk. The work might be disputed and the client might be slow, but the receivable is a contractual claim that can be sold. Litigation finance monetises a lawsuit before it pays, an asset with no liquidation value at all, just a probability and a judgment. Media royalty deals securitise income from songs not yet streamed. Revenue-based financing lends against the recurring subscription line of a software business that owns almost nothing you could touch.

In every case the breakthrough was not repossession. It was legibility plus a contract that isolated the economic right from the people and the process. Litigation finance never learned to foreclose on a lawyer. It learned to write a claim on the outcome.

The work is now legible enough to write an isolable claim against. That is the real unlock for services, and AI is what makes it possible at scale.

Services financing is not unprecedented. It is the next expansion of finance into a cash flow that used to be too tangled to price.

The Destination Is Slower, and the On-Ramp Is Bigger

The second mistake was speed. We jumped from underwriting to securitisation in a single paragraph, the way the slide always does. Reality moves through stages. The early stages are where the money sits for the next decade.

So we concede the title ran ahead of the mechanics. Before professional services is an asset class, it becomes underwritable working capital. That is the bigger near-term prize anyway. Pipeline-backed lending. Credit limits that move with delivery health. Delivery risk priced into the cost of capital. Insurance against scope failure. A secondary market in engagement claims. None of that needs a single ABS to exist, and all of it changes who can fund a services firm and at what price.

Calling it a brand-new asset class oversells the maturity. It is more honest, and more useful, to say professional services is acquiring the characteristics of a financeable asset pool, and that it arrives first as a subclass of private credit, not as a wing of its own.

What Survives the Critique

Strip out the overreach and the core claim is still standing, and still under-priced. The quality of a firm's scope and delivery data is becoming an input to its cost of capital. The firm with machine-readable, evidence-backed, assignable engagements will borrow more cheaply than the firm whose work lives in a partner's head and a slide nobody can audit. That gap did not exist five years ago. It is opening now.

The underwriting spine of professional services is real. We drew it one link short and one decade too fast.

Observability gets you noticed. Assignability gets you funded. The firms that understand the difference are the ones the capital will reach first.