Second leg — AI data licensing (finder / advisor income)

Added Oct 13, 2026. Background: Chad's own research artifact "AI Data Licensing: Mercor & the Market" (Oct 7, 2026), pasted in full to Muse Oct 13. Key facts below are from that artifact, attributed to its cited sources; treat figures as reported, not independently verified by Muse.

The market in one paragraph

Frontier labs (OpenAI, Anthropic, Google, Meta, Nvidia) need real company workflows — Slack, email, tickets, code — to train agents and build RL environments, and they mostly buy through brokers. Mercor alone booked $614M gross revenue in H1 2026, >90% from model companies, at a $2B+ run rate. A single high-fidelity Slack replica runs ~$300k (Epoch AI). Datasets get resold to multiple labs unless the contract forbids it; some brokers (Replay, Troveo, Protege) pay sellers a revenue share per resale, Mercor's published model is a single payment.

Three ways Chad gets paid (from the artifact, §3 and §5)

1. Public referral bounties — start now. Scale AI: up to $50k per company that signs (40+ employees preferred). micro1: up to $50k (30+ employees, mostly English data, US prioritized). Mercor: amount currently unstated; a tracker showed up to $100k advertised Jul–Sep 2026. Pays on signing/go-live/milestones depending on the program; no tail, no protection if the seller finds the buyer directly.
2. Negotiated channel/finder agreement. Show deal flow first, then negotiate a percentage of contract value, renewals, a 12-month tail, and non-circumvention backed by a written register of introductions. Target buyers: Mercor (Chad already has Garrett/Owen relationships), SimpleClosure, Replay, Troveo, Halluminate.
3. Seller-side advisor. Represent the seller, run competitive bids across buyers, take a success fee (the Data Licensing Advisors model — the closest existing analogue to this role). Higher effort, carries more compliance burden; never take fees from both sides without disclosure.

The math against $20k/month

This leg is lumpy, not monthly — it averages out. One signed Scale/micro1 referral ≈ up to $50k ≈ 2.5 months of target. A 10% success fee on one $200k seller deal = $20k = a full month. Two or three closes a quarter across referral + advisor models covers the target on its own; combined with the physical-goods leg's steady turns, the two legs de-risk each other.

Chad's specific edge (why him, from the artifact's own criteria)

• He is a live Mercor data partner ($175k deal in progress) — he has already passed the diligence sellers worry about and has named contacts on the buyer side.
• Former CPA + agency owner: credible with the exact channels that surface sellers (accountants, startup lawyers, wind-down advisors, VC platform teams) and fluent in the client-data/PII screening buyers want pre-cleared.
• Location: two direct buyers are local — Appen (US HQ Kirkland) and Defined.ai (Seattle-founded, claims $600k+ average partner revenue).
• Seller profile buyers want: 30–200 employees, 3–5+ years of English-language records, linked systems (Jira tied to commits), an exec who can sign; best timing is winding down or just-acquired companies. Poor fits: firms whose data is mostly client-owned — which is exactly the screening Chad can do better than a cold lead list.

Cautions carried from the artifact

• Chad's own Ambaum data: the Mercor DLA he is negotiating is exclusive/perpetual in its current draft (his Oct 12 reply to Owen already contests the exclusivity point) — re-selling or re-licensing his own dataset elsewhere depends on how that lands. The finder/advisor income does NOT depend on his own data, which is the point of this leg.
• Platform terms are a real screen: Shopify Partner Agreement §9.14–9.15 bars sharing merchant data and AI training on it without consent; Slack API terms bar LLM training on API data; buyers' extraction has to work within these. Pre-screening sellers for this is part of the value Chad sells.
• Finder-fee law: WA's sales-rep commission statute likely does not cover data deals, so protection must be contractual; NY requires finder compensation in writing (GOL 5-701). Lawyer drafts the template (artifact §5 checklist: percentage + floor, renewals, trigger, ≤30-day payment, 12-month tail, non-circumvention, audit rights).

First 30 days (artifact §5, adopted)

1. Register for Scale, micro1, and Mercor referral programs; save current terms.
2. Build a 100-company target-seller list (30–200 employees, target verticals: legal, finance, insurance, professional services, manufacturing, real estate, logistics, healthcare-with-de-id-path).
3. Ask Mercor (via existing contacts) about a channel agreement.
4. Pitch SimpleClosure, Replay, Troveo, Halluminate on the same terms.
5. Lawyer drafts the finder-agreement template.
6. One-page seller explainer: payouts, process, client-data limits, privacy handling.