The Middleman-Signal Disintermediation Play
When middlemen and end-customers both come to you directly, that's the signal to build the business yourself
- Difficulty
- Advanced
- Time to result
- ~months to results
- Steps
- 4
- Confidence
- 85%
A pattern for discovering and validating an AI-era adjacent business built on an asset you already own. The trigger is two simultaneous inbound signals — middlemen wanting to license your asset, and end-customers reaching past those middlemen to go direct to you. Garrett used it to recognize, validate, and launch Handshake AI on top of Handshake's existing expert network.
Origin
Described by Garrett Lord as how Handshake AI was discovered, starting over Christmas/New Year 2024-25.
Core principles
- 01Watch for middlemen monetizing your asset — it proves demand and that you're sitting on scarce supply
- 02End-customers going direct to you, trying to cut out the middleman, is the clearest signal to disintermediate and own the relationship
- 03A mature core business's resources (cash, brand, relationships) can be spent to accelerate the learning curve on a new bet
- 04Validate fast and cheap with expert-network calls before committing capital and headcount
How to run it
- 1
Notice the inbound demand for your asset
Track when third parties start asking to access or license the asset you already own (Handshake's PhDs and master's students being recruited by middleman data companies).
Pro tip Route the early demand through partners first — the friction users report on those platforms is your product spec.
Watch out If you only monetize as a passive referrer, you leave the margin and the customer relationship on the table.
- 2
Watch for customers going direct
Treat end-customers (frontier labs) reaching out to bypass the middleman as the decisive signal that you should build the direct business.
Pro tip When both middlemen and end-buyers converge on you, you can improve the experience for the audience, the buyers, and your existing network simultaneously — a rare win-win-win.
- 3
Validate cheaply with expert calls
Before committing, hire expert-network firms and run many calls with practitioners to compress the learning curve.
Pro tip Use the core business's balance sheet to fund fast learning — a $200M-ARR base can afford to buy speed most startups can't.
Watch out Skipping structured validation risks scaling too fast and eroding the trust with buyers that the whole business depends on.
- 4
Commit through personal hustle, then land the anchor customer
Chase the leads personally (Garrett flew around over the holidays), build the team, and win one marquee customer before broadening.
Pro tip Deliver flawlessly to a single top customer before expanding to others — protect trust over growth speed.
Watch out Growing too fast can erode the trust you've built with frontier buyers; deliberately say no to a lot of things early.
In the wild
Middleman data companies asked to recruit Handshake's PhDs and master's students; then frontier labs reached out directly to cut out those middlemen. Garrett spent the holidays flying to leads, hired three expert firms (Alphasights, GLG-type) for research calls, built a team from the human-data world, and landed 'arguably the number one lab' about five months in.
→ Launched January, monetized within months, reached $50M ARR in four months working with seven frontier labs.
Common mistakes
Staying a passive referrer
Handshake was already making tens of millions sending experts to other platforms; treating that as the ceiling would have missed the far larger direct business the inbound signals pointed to.
Expanding before nailing one customer
Chasing many buyers early risks quality slips that erode hard-won trust — deliver to one anchor lab first.
Is it for you?
Best for
Leaders of an established business receiving inbound interest in an underused asset, deciding whether to disintermediate and build a direct business
Not ideal for
Companies without a genuinely scarce owned asset, or without the resources to fund fast validation and a dedicated team
From the transcript
“we started to see all the what I would call like middleman companies reaching out to us saying, can we recruit your PhDs and master…”
“they started to reach out to us and sort of go direct”
“I hired like three expert firms, alpha in the alpha sites and like GLG and started doing a bunch of calls with the latest researchers”
“making sure that we don't grow too fast as to erode the trust that we've built up with these frontier labs”
From the episode
Inside the expert network training every frontier AI model
Garrett Lord (Handshake CEO)