What the founder is really asking you to carry
"Promise me there's a buyer" is rarely a question about the AI market. It usually means one of three things: is this worth my team's time, will I look foolish in front of my board, and will I end up handing over information for nothing. A promise would move that risk from the founder to you, and you cannot carry it, because you decide nothing in SourceX's review.
What you can do is make the test cheap. SourceX's first step is a few questions with no files, ending in an indicative licensing estimate. That costs a founder far less than the certainty they asked for, and it produces evidence about their own company instead of reassurance from you.
Four levels of demand evidence
Every demand sentence rests on one of four kinds of evidence. The level decides what you may say, and who may say it.
| Level | What it is | Who holds the evidence | The strongest sentence it supports | What it cannot tell the founder |
|---|---|---|---|---|
| 1. Market announcement | A public statement of another company's plan or budget, such as micro1's October 9, 2026 announcement of a commitment to spend $1 billion on enterprise data acquisition over 12 months | The announcing company | "Some companies have publicly committed budgets to business data; that is a signal about the market." | Anything about this company or SourceX's programs; a plan is not spending |
| 2. Published fit | SourceX's screening profile, which it calls screening guidance rather than a promise of eligibility, with requirements that vary by current data needs | SourceX's public pages | "On SourceX's published criteria you look like a candidate; SourceX decides relevance in review." | Whether the records match a current program |
| 3. SourceX's assessment | The company's own application: a few questions, no files, an indicative estimate, then qualification and an indicative range | The company and SourceX | Said by the company, not you: "SourceX qualified us and gave an indicative range." | Whether anyone commits; qualifying does not guarantee an offer |
| 4. Signed and committed | A Dataset Deal Confirmation the company signs with SourceX, fixing the dataset and payout; SourceX then seeks commitment and signs its own onward agreement before processing | The signed documents | "A specific dataset has a fixed payout in a signed Confirmation." | Nothing before it; and payment still waits for Closing |
Your own voice stops at level 2. At levels 3 and 4 the company speaks about its own status, and partners never receive confidential buyer information.
Method note. The levels were defined from SourceX's process page and the program terms, checked 2026-10-27, by asking who holds the evidence behind a sentence. They grade evidence, not probability. SourceX has not published how many companies move from one level to the next, and this page does not estimate it. Two longer treatments sit nearby: the eight-stage ladder from fit check to buyer acceptance and the three-tier demand checklist.
Rewrite the sentence, not the founder
Most overpromising comes from a sentence that claims a higher level than the speaker holds. Lower the claim to match the evidence.
| Tempting sentence | Level it claims | Level you hold | Honest rewrite |
|---|---|---|---|
| "AI companies are desperate for data like yours." | 3 or 4 | 1 | "Some AI data companies have announced large budgets. Whether your records fit is SourceX's call." |
| "There's a buyer waiting for records like yours." | 4 | 1 or 2 | "Nobody has looked at your records yet; the assessment is how you'd find out." |
| "You'd get at least six figures." | 4 | None | "Any figure would be SourceX's indicative estimate. I won't guess one." |
| "You qualify, so it's basically done." | 4 | 3 | "Qualifying is a start. Offers come later, and some never come." |
| "Once you sign, the money's yours." | Beyond 4 | 4 | "Payment follows Closing, and data that doesn't close earns nothing." |
| "Demand will only grow from here." | A forecast | 1 | "I won't forecast the market. I can tell you what's been announced, with dates." |
| "I'll make sure SourceX takes you." | None | None | "I don't decide anything in review." |
SourceX's terms apply the same discipline to partners: estimates in the portal or calculator are projections, not promises. The page on whether AI data demand is a bubble covers the market question for a founder who wants to argue it.
What is known, conditional or unreviewed for this company
- Known now: how SourceX's process runs, that the first step needs no files, that the company signs nothing binding until a Confirmation, and that it can stop at any point before then.
- Conditional on facts the founder holds: fit. The company needs 50 or more full-time employees at its peak (contractors excluded), a documented operating history, the right to license its records and a CEO who wants the conversation. Companies in the U.S., Canada and the EU qualify, as do companies in the UK, Australia, New Zealand, India, Singapore, South Africa, Latin America, the UAE and Saudi Arabia when at least 70% of the data is in English. Records already licensed for AI training are out, because the supplier confirms they never were.
- Unreviewed until SourceX looks: relevance to a current program, rights under customer contracts, any price, and whether an Acquisition Partner commits.
A founder who asked for a promise (Illustrative)
Illustrative: an invented administrator and conversation. The founder-CEO of a 260-person third-party claims administrator in a growth fund's portfolio tells the operating partner: "If you can tell me a buyer wants our claim notes, I'll spend time on this. Otherwise, no."
"I can't tell you that, and I'd be wary of anyone who did. Here's what I can say. Some data companies have publicly announced big budgets for business records; that's a market signal, not a buyer for you. On what you've told me, 260 people and twelve years of claim notes in one system, you look like a candidate against SourceX's published criteria, but relevance is SourceX's decision. The way to find out is its assessment: a few questions, no files, and an indicative estimate at the end. Nothing is fixed until you sign a Confirmation, and nothing is paid unless the data closes. If twenty minutes isn't worth that much uncertainty, that's a fair reason to pass."
The reply climbs the levels in order and stops where the partner's evidence stops. If the founder applies, the next facts come from SourceX to the company. If the founder passes, the partner records "declined: wanted certainty first" and leaves it there.
One voluntary next step, and a clean way out
Offer exactly one step: the company's own assessment, or the company fit checker for a preliminary look on the founder's own time. Any question that needs company-specific demand goes to SourceX through that application, never through you on the founder's behalf. The CEO decides; a CFO or board member can advise but does not authorize. If the answer is no, a short close works better than a second pitch, and the guide to a respectful close has the wording.
Where this approach could be wrong
- Market signals overstate. micro1's commitment is one company's stated plan, not money spent, and it says nothing about demand for any given company's records.
- Synthetic data could narrow demand. NVIDIA states that over 98% of the data used in the alignment process for its Nemotron-4 340B model was synthetically generated. If more developers substitute synthetic data, demand for human-generated records may shrink for some uses.
- Programs move. SourceX applies the criteria of specific programs, and a company that does not fit one may fit another later; the reverse can also happen.
- The commitment step can stall. SourceX's process calls its 7 to 14 day window for commercial feedback a target, not a promise, and positive interest is not yet a purchase.
- Honesty costs introductions. Some founders will pass when no one promises demand. That is the price of a claim you can defend.
Next step
Refer a company. No cap on rewards. Default reward on a company's first qualifying transaction: the greater of $50,000 plus 5% and 15% of Actual Supplier Payout, with no cap; further qualifying transactions first agreed within 12 months of the first Closing pay 15%. Eligibility and payment conditions apply, including the USD 100,000 Minimum Transaction, and a signed Schedule A or addendum can differ. A lead, meeting or signature alone does not trigger payment.
If a founder accepts a graded answer and wants to run the assessment, register for a partner link, sign the Channel Partner Agreement, check the company against who qualifies and read the program terms; other founder objections are covered in the guides hub.
Readers weighing SourceX as an investment rather than a referral program face a different question, and can request investor materials through the SourceX investor overview.
This article is for information only and is not an offer to sell or a solicitation of an offer to buy securities. Any offering by SourceX is made only to verified accredited investors through its offering documents.