Mercor Company Data: What It Buys, Pays and Asks

Yes. Mercor buys companies' operating records and licenses them to AI labs: it connects to the tools a company runs on, extracts and anonymises the records with its own automated pipeline, at its own cost, and pays by wire on delivery of the anonymised dataset, typically 2 to 4 weeks after signing. Its data page publishes no figure, only that compensation is based on data volume, the tools you connect, and the depth of your data, but its referral listings do: 25 listings, one per kind of company, each stating what such a company earns, from $10,000 to $1,500,000. Mercor is also the alternate bidder for Spirit Airlines' records, at $7.5M, which puts one of its bids for a whole company's archive on a court record. What it does not publish is the licence it takes: how long, whether exclusive, and whether it may pass your data on.

This guide is for the founder, CEO, CFO or counsel of a company that has heard Mercor pays for company data, or has had an email from it, and wants to know what it buys, what it pays a company like theirs, whether they qualify and what they would be agreeing to. It reads Mercor's own pages, its listings and the court record of the Spirit sale, in the order a company meets them: what Mercor is and why it buys, what it takes and the tools it connects to, its listings by kind of company, who it takes, its pipeline and what happens to personal data, how and when it pays, the licence, the Spirit bid, and how to apply or be introduced. It ends with a comparison against Turing's Project Lazarus, micro1 and Scale AI, and a worked example.

AIData.Broker is in Mercor's referral programme, as it is in Handshake AI's, Turing's, micro1's and Scale AI's. We introduce a holder to every buyer whose published rules take its data, free for the holder, because the buyer pays us its published referral payout out of its own money: for Mercor, an amount set on each listing, the most for any one company being $85,000, and the holder sees the most any buyer may pay us before we introduce it. Where a sale run by us earns more, and only for holders based in the United States, the United Kingdom and Canada, we act as the holder's agent under a mandate and the holder pays our fee only if the sale closes; every fitting partner is introduced then as well. Either way the holder itself licenses its data to the buyer, and we do not buy it, take ownership of it, or license or sell it in our own name. Every term below is Mercor's own published wording, a court record or named press, cited to its record; where a line is our reading, it says so.

This guide is research from public sources, not legal, tax or investment advice.

Contents

1. What Mercor is, and why it buys company data
2. What Mercor buys, and the tools it connects to
3. The listings: what each kind of company earns
4. Who it takes
5. The pipeline and what happens to personal data
6. How and when it pays
7. The licence: what is published and what is not
8. The Spirit Airlines bid
9. Applying directly or being introduced
10. Mercor beside Turing, micro1 and Scale AI
11. Worked example: a litigation firm
12. Questions people ask
13. What this means for you

What Mercor is, and why it buys company data

Mercor started as a hiring platform and became one of the largest suppliers of human expertise to AI labs: it builds benchmarks, RL environments and large-scale human datasets for labs, through a marketplace of experts it says numbers nearly five million. It raised $350 million at a $10 billion valuation in October 2025, was reported in July 2026 to be in talks to raise at $20 billion (no closed round has been found since), and was reported at over $2 billion in gross revenue run rate in June 2026. Most of that money is paid out to the experts who do the work (by our estimate Mercor keeps 30% to 40% of each dollar a lab pays for it), so it is a measure of the business's size, not of what it spends on company data.

Mercor buys company records for the same reason it hires experts: labs want models that can do real work inside companies. Its data page puts it this way: "Frontier AI models are increasingly being trained to work inside companies, navigating real tools, workflows, and decisions", and public data and synthetic datasets cannot replicate that ( data page). Its acquisitions point the same way. In July 2026 Mercor agreed to buy Deeptune, which builds reinforcement-learning environments and says it has made 100s of gyms that mimic popular software like Slack and Salesforce; earlier in 2026 it bought Sepal AI, a data research company. It is also hiring a research engineer to capture large enterprises' data and turn it into RL environments and training datasets for frontier labs. A company's real Slack threads, tickets and documents are the material such a simulated workplace is built from.

Mercor is a principal: it licenses your records for its own account and licenses the result to labs, and its own catalogue offers every dataset to labs for licensing. It also reaches out to companies itself: its data page tells introducers that once they name a company, "We handle the outreach", and AI Weekly reported a note from Mercor to Warmly, after Warmly agreed to be acquired by HubSpot, offering to buy or license Warmly's code base and the records of its tasks. Warmly received four offers of up to $300K for its code and communications and turned them all down, which is what a holder that does not need to sell can do. No holder has yet confirmed in public a sale of company records to Mercor; its one recorded purchase attempt is the Spirit bid below.

What Mercor buys, and the tools it connects to

Mercor buys workflow data: the messages, documents, tickets, accounts and code a business produces by working, taken from the software it already uses. Its data page lists the connectors it shows first, each with what it takes: Slack ("Workspace messages, channels, and threads"), Microsoft Teams, Gmail ("Email messages and metadata"), Confluence, Salesforce ( accounts, contacts, opportunities, and reports), Outlook, GitHub ( repositories, commits, PRs, and issues), Google Drive and Notion ("Pages, databases, and workspace content"). Behind those it says you can choose from 100+ supported tools and link your accounts. Its listings add the specialist systems each kind of company runs on, from Procore and Revit in construction to Epic and athenahealth in medicine and Relativity and iManage in law.

The more systems you connect, the more it pays, in its own words: " The more tools you share, the higher your payout will be." That is a real lever, and other buyers' estimators show the same effect: Ooak's values a mid-sized company with six tools at several times what it shows with three ( its estimator). It is also a lever on your risk, because every connected system brings its own rights questions: whose messages are in Slack, which customers are in Salesforce, which contracts forbid sharing what is in Drive, and what the software's own terms allow. Read our pages for the systems you would connect, such as Slack, Microsoft 365, Google Workspace, Salesforce and GitHub, before you connect any of them.

The eight industries on its data page

The data page names eight target industries, each with the records it wants ( data page): legal ( contracts, case files, discovery docs), healthcare (clinical notes, claims, intake forms), finance and insurance ( statements, underwriting files, claims), professional services ( client decks, SOWs, project files), manufacturing ( work orders, QA logs, spec sheets), retail and e-commerce ( order records, inventory logs, listings), real estate (listings, leases, inspection reports) and logistics ( shipping manifests, dispatch logs, invoices). "If your company operates in one of these spaces," it says, "there's a good chance your workflow data is exactly what labs need."

The list is a guide to demand, not a rule. Its listings cover more than these eight (game studios, cybersecurity, semiconductors, publishers), and a company outside them is not turned away by any published word. What the list does show is the kind of record Mercor values: documents that carry a decision and its outcome (a claim and its adjudication, a work order and its inspection, a case file and its result), which is what an AI model can learn a job from.

The listings: what each kind of company earns

Mercor's referral page, written for introducers, is where it publishes prices. Since 17 October 2026 it has listed one listing per kind of company, each with a range of what such a company earns, the size and place it asks for, the records it requires and those it would also like. There are 25 today. A listing prices a kind of company; where one company lands within its range is set after Mercor reads its records, and each listing names what moves it ("Depends on matter size and completeness", for a law firm). The tables below are Mercor's own figures as its listings showed them on 6 November 2026, grouped by kind of company; each line links to its record, where the listing's words are kept.

Health and life sciences

The health listings carry the most conditions. The medical-practice listing asks for an EHR that holds free-text physician notes, not only structured fields, and at least 2,500 patients or 10 clinicians; the imaging listing for DICOM studies paired with signed reports; the digital-health listing for recorded visits with consent that covers secondary use. Each of those is a health-privacy question before it is a commercial one (in the US, whether the records meet HIPAA's de-identification standard): read HIPAA de-identification for AI training and the pages for medical practices, imaging centres and digital health companies.

Law, accounting and finance

Two of these ask for records that carry legal duties of their own. The accounting listing is the only one that requires the company to be in the United States, and it wants tax returns with workpapers across several entity types and several filing seasons, which brings in the federal rule on tax preparers' use and disclosure of return information: for a firm that prepares returns, licensing them without an exception or each client's specific consent is a federal crime, not only a breach of contract. The law-firm listing wants matters with 500+ documents and client consent, or documents that can be fully anonymized. The pages for law firms, accounting and tax firms and investment firms set out the rights questions for each.

Engineering, industry and construction

These listings are specific about form. The hardware listing wants native assemblies with feature trees, not exported meshes, and rules out customer-owned designs and part catalogs; the construction listing wants drawings, specs, RFIs, submittals, change orders and daily logs from completed projects, tied to site imagery. The guide to selling CAD files and the pages for industrial and engineering firms and construction and trades go further.

Software, data and security

The software listing says it pays more for large codebases or exclusive terms, the clearest sign on any of its pages that exclusivity is priced. A company whose code is most of what it holds should also read the guide to selling a codebase to AI companies and the page for software companies.

Media, design and services

How to read the ranges

Three things about these figures matter more than any single row. First, they are wide on purpose: most start in the tens of thousands of dollars and end many times higher, because Mercor prices by volume and depth once it has read the records, and the top of a range is the most such a company can earn, not what a typical one receives. Second, they are the ranges Mercor shows introducers; the data page a company sees shows none, so a company that applies directly will hear the figure only after scoping. Third, they are not exclusive of other buyers' figures: a construction company here may also fit buyers that publish their own ranges, and the buyer board and our price board show the rest. Compare them with an independent valuation of what you actually hold, and with what company data is worth.

Who it takes

Mercor's rules live in its listings, one kind of company at a time, rather than in one threshold. Each listing states the size it wants, where it takes companies from and the records the company must hold. Its data page adds no general rule beyond the eight industries.

Size. The smallest listings start at five people (law firms, accounting firms, investment firms, AI product companies, hedge funds, video studios); the imaging, payments and R&D listings take companies of any size; construction and SaaS companies start at 50; call centres are counted in agents. So Mercor takes smaller companies than most buyers of company records, several of which start at 20 or 30 staff. The size is a listing's rule, though, so a company that falls between two listings, or fits none, should ask rather than assume.

Place. 18 of its listings say "Anywhere", the game-studio listing adding that international studios are welcome; the law-firm, imaging and medical listings prefer the United States (the medical one adds that other countries are considered); the accounting listing requires it; and the call-centre listing names its markets. That puts Mercor among the 7 buyers of company records whose own pages say they take companies from anywhere or work with companies abroad. Its payment terms state no currency or payout country, and its general terms of service ask users to confirm they are not on a US sanctions list and not in a country under US sanctions. A company abroad should read selling data from outside the US and US withholding tax on data licence payments first.

Records. Each listing names records a company must hold, and they are demanding in the same way: complete sets, with history, that the company owns. The law-firm listing wants closed matters, not single filings; the hardware listing wants designs the company owns, not customer IP; the video listing wants original footage the company owns, with appearance releases, and 100+ hours of video. A company that cannot show it owns the records, or has the right to license them, does not fit any listing, whatever its size. Mercor's intake asks you to confirm the materials are ones you own or are authorized to share without restrictions.

Kinds of company it does not list. Its listings have none for a business that runs on bookings, such as a salon, a clinic front desk, a restaurant or a home-service company, and none for staffing and recruitment. Such a company is not ruled out by any word on its data page, but has no published figure to work from; the guides to booking and scheduling data and recruitment data name the buyers that take those records.

The pipeline and what happens to personal data

Mercor's process has four steps on its data page: scope (choose the tools to share), connect ( link your accounts among its supported tools), process (its platform "extracts and de-identifies data into an asset") and payout. It says that no data leaves your systems until you approve it, that extraction and anonymisation are fully automated and paid for by Mercor, and that no AI lab buyer or third-party vendor ever sees your raw data; its engineers oversee the pipeline but do not manually review your data, and all access is logged. The same page lists encryption in transit, credentials held per session and never shared between connections, SOC 2 Type II, and encryption keys managed in AWS KMS ( data page).

That is a stronger set of statements than most buyers publish, and two parts of it deserve a question each. The first is where processing runs. Connecting your accounts gives Mercor's platform access to them, and "no data leaves your systems until you approve it" leaves open what the platform reads before you approve, and where the raw copy sits while it is processed. Scale AI, by contrast, says it anonymises inside your own systems before anything is extracted. Ask Mercor which systems its connector reads, where the extraction and de-identification run, how long the raw copy is kept and how it is deleted. The second is assurance. Mercor posted an update on a security incident in 2026, saying only a limited subset of its experts had sensitive information affected. Before any record moves, ask for its current SOC 2 Type II report and, for patient records, a business associate agreement.

The legal duty stays with you. Mercor's de-identification protects Mercor's use of the data; it does not decide whether you may hand the records over. Customer records raise what your privacy policy promised ( is it legal to sell customer data); staff messages raise your employees' expectations and, in some places, their unions ( employee emails and chats in AI training); client files raise your contracts and, for professionals, your duty of confidentiality. And who controls de-identification is itself priced, as Mercor's own Spirit bids show below. Our de-identification standard and tool show what a holder can do before any buyer sees a file.

How and when it pays

Mercor's payment term is the most concrete of the large buyers'. Its FAQ: "Payment is issued via wire transfer upon delivery of the anonymized dataset to the buyer. This typically occurs within 2-4 weeks of the signed agreement, depending on data extraction and processing time." And: "Mercor covers all costs associated with data extraction, anonymization, and transfer. You receive compensation with no upfront costs, setup fees, or deductions from the payment." Other buyers have done the same: Google and micro1 bore all de-identification costs in the Spirit sale with no price reduction.

The same page words the trigger differently in its process steps: " Get paid as soon as processing is complete." In a smooth deal the two events are close. In a slow one they are not, and the FAQ's wording raises a further question: "delivery of the anonymized dataset to the buyer" can be read as delivery to a lab that licenses it from Mercor, which would make your payment wait on Mercor's own sale. Our reading is that the plain intent is payment once your dataset is processed and delivered, within weeks of signing, but a web page that is not in the signed agreement may not bind, so ask which event the agreement names, how many days after it the wire is sent, and whether payment depends on any lab licensing the data. Keep a dated copy of the page you relied on.

Against the other large programmes, Mercor's clock is short and its trigger close to the holder's control. micro1 pays only after it accepts the data, with 30 business days to reject it; Handshake AI pays on delivery milestones set in each agreement; Turing pays from each sale of your record to a lab. For every buyer's payment trigger and day count side by side, read when AI data buyers actually pay.

The licence: what is published and what is not

Mercor's data page says only: " We license it, package it, and pay you for it." Nothing published says how long the licence runs, whether it is exclusive, whether Mercor may sublicense (pass the data or the rights on), which labs may receive it, or what happens to trained models if the licence ends. Its catalogue offers every dataset to labs for licensing, so the working assumption should be that Mercor may license your data to more than one lab, and the agreement should say so, or say otherwise, in terms.

The nearest published terms are Mercor's general website terms, which cover anything a user provides through its service. They take a licence over that content that is irrevocable, perpetual, transferable and sublicensable through multiple tiers, and they also say that separate agreements you sign with Mercor govern ownership and override the website terms where they conflict. Our reading: those terms were written for experts and visitors, not for a company licensing its records, but a company connecting its tools through Mercor's portal is a user of that service. So ask that the data licence agreement state plainly that it alone governs your data, and that the website terms' licence does not reach it.

Two of Mercor's listings show that it prices the terms. The software listing pays more for large codebases or exclusive terms, and the hedge-fund listing says its price depends on history and exclusivity. On the market's evidence an exclusive licence fetches 2.0 to 5.8 times a non-exclusive one, so a holder asked for exclusivity without a separate price for it is giving away its largest lever. Disclosed AI licences run two to three years or at least three years, and licensing counsel recommends pricing a perpetual right to models already trained separately from a finite licence to the data. The licence terms page, the guide to exclusive and non-exclusive licences and the offer check help you decide where to stand.

The Spirit Airlines bid

Spirit Airlines' bankruptcy produced the one public auction of a company's internal records, and Mercor was in it. The estate offered de-identified business data: email, Teams, files, tickets and code, and the competitive bidders agreed an asset schedule that excluded personal information and set passengers' data aside. Google opened at $5M. Mercor's first overbid was $5.2 million on Google's terms, where an independent party de-identifies the records, or $7 million if it could do the de-identification with its own tools; it later offered $10 million on in-house terms. The debtors chose the lower, independently scrubbed bids ( the auction record). Google won at $10M, and Mercor's bid was designated the alternate bid, at $7.5M.

Three things follow for a company dealing with Mercor. First, Mercor valued the right to de-identify with its own tools at 35% more for the same records. That right is the default in its ordinary programme, where its own pipeline does the work, so a holder that wants an independent scrub is asking for something Mercor has shown it prices. Second, the estate did not simply take the highest number: a later bid from micro1 of $12.5M has not displaced Google, because in a privacy-sensitive sale, structure beats a higher late price. Third, the sale is still not approved: the debtors moved the approval hearing from 6 November to 17 November 2026, while they negotiate with unions over employee information. An alternate bidder buys only if the winning sale falls through, so Mercor's bid remains a bid. For the full case, read selling customer data in bankruptcy.

Applying directly or being introduced

There are two ways to reach Mercor's data team. Both lead to the same scoping, the same pipeline and an agreement you sign with Mercor.

Applying directly

The data page has a "Get a quote" button and a "Check if you qualify" link beside its industries ( data page). Before you use either, decide which tools you would connect and which you would not, which records inside them are out of scope (named clients, privileged files, health data, a director's mailbox), how many years each holds, and who in the company can sign. Mercor writes the agreement; you decide the scope. A company that applies itself saves the buyer whatever it sets aside for an introducer, and buyers budget a share of a deal, or a fixed amount, for whoever brings them a company, so a holder that comes direct can ask for that budget in its price, though nothing Mercor publishes says it agrees to that.

Being introduced

Mercor's referral programme is open to anyone with a Mercor Data partner account: "Do you know someone who runs a company? Refer them to Mercor. When they become a paid partner, you get rewarded too" ( its data page). It says the steps are a short referral, then "We handle the outreach, agreement, and data pipeline from start to finish", then the bonus once the company goes live. Its referral terms count only the first valid referral of a company, do not count a company already working with Mercor, confirm the introducer's payout when the company signs its data licence and pay it after the first delivery, and allow each introducer up to five companies awaiting review at a time. The company pays nothing for being introduced and signs its own agreement with Mercor.

This is how an introduction through us works. Value your data first: if one of Mercor's listings takes your company, the valuation says so and offers one button, Introduce me. Pressing it is your consent to the introductions it names, and we file each one in the partner's own referral form within one business day. Mercor then writes to you itself, by email, with us copied; that email is the introduction, and answering it is your next step. From there, scoping, the agreement and the payment are between you and Mercor. Because Mercor limits how many introductions each introducer may have awaiting its review, its places are scarce, and the valuation offers Mercor only while a place is free for a company like yours; when none is, you can apply to it directly, and the same press still introduces you to every other partner whose rules take your data. For what each route costs, read going direct or through us.

Mercor beside Turing, micro1 and Scale AI

A company weighing Mercor usually weighs three other programmes run by large suppliers of human data to labs. Handshake AI, the fifth, publishes a single range for every kind of company and has its own guide, as do Turing's Project Lazarus and micro1.

Turing pays differently from Mercor in kind, not just in amount. It prepares a company's record, licenses it to labs and pays a revenue share on every licence sold, with nothing upfront; if nothing sells you owe nothing. Its licences are non-exclusive, so the same record can go to more than one lab, and you keep ownership. Against Mercor it trades one payment within weeks for the chance of several later ones, and it gives no figure until it has reviewed a sample within 5 business days.

micro1 publishes tiers rather than listings, and the broadest licence of the four: for ever, worldwide, paid in full once and irrevocable, which it may grant on to others, who may grant it on again, and since 30 October 2026 it owns everything made from the data. It wants English-language records and ranks US companies first, then other Western markets, so a company abroad has a clearer published path with Mercor. It asks you to promise you have not already licensed the same data on similar terms, which matters if you license to Mercor first.

Scale AI is the one to compare on handling: it anonymises inside your systems before extraction and releases data package by package, each with your sign-off, where Mercor's pipeline extracts first. It publishes an illustrative $10K to $1M+ and says it is still defining how these deals are priced; it prefers 40+ employees, where several of Mercor's listings start at five.

Read side by side, Mercor is the most specific about price and the most open about place and size, and among the fastest to pay; it is as silent as Handshake on the licence it takes. Its figures are therefore the easiest to compare and its terms the hardest. The guide comparing every company-data buyer programme covers a dozen more buyers, and the buyer match tool checks your company against each one's published rules.

Worked example: a litigation firm

Example. A litigation firm in Columbus, Ohio, has 40 attorneys and staff and twelve years of closed matters in iManage and Relativity, with its email in Microsoft 365. Mercor writes to it. The managing partner wants to know what Mercor would pay, what it would take and what the firm must settle first.

Does it fit? Mercor's law-firm listing asks for 5 to 300 attorneys and staff, United States preferred, and 5+ years of closed matters, and requires complete files from closed matters or completed transactions, matters with 500+ documents, and client consent, or documents that can be fully anonymized. It would also like custodial email and chat from discovery, e-discovery exports, contract repositories and transaction data rooms. The firm fits on size, place and history; whether it fits on records depends on how many of its matters are complete and how many clients would consent.

What might it earn? The listing's range is $50k to $1.2M, and its own note says the figure "Depends on matter size and completeness". A firm whose matters are mostly small, or whose files are incomplete, should expect the lower part; one with large, complete litigation files and data rooms has the material the top of the range is for. The firm's own valuation against published prices is the reading to set beside Mercor's number once it arrives.

What must it settle first? Client files are the client's confidences before they are the firm's asset. The firm decides which matters it may license at all (no active matters, nothing a client has not released), whether it will seek consent matter by matter or rely on anonymisation, and what anonymisation means for a file whose facts identify the client even without names. It checks its engagement letters and its software terms for iManage, Relativity and Microsoft 365. Privileged material has to come out before anything moves: the federal rule that rescues an inadvertent disclosure of privileged material ( Fed. R. Evid. 502(b)) does not cover a deliberate commercial transfer, and the Spirit sale agreement shows the standard shape of the exclusion, where privileged material is excluded and anything sent by mistake must be returned or destroyed. The page for law firms sets out the rest.

What does it ask Mercor? For the full agreement before connecting anything; what its connector would read in iManage and Microsoft 365 before the firm approves a scope; where the extraction and de-identification run and who checks them; the licence's term, exclusivity and sublicensing; whether "delivery to the buyer" in the payment clause means delivery to Mercor or to a lab; and its current SOC 2 Type II report. It also puts the same matters before other buyers: Handshake AI ( 20+ full-time staff, 3+ years), micro1 ( 30+ staff) and Scale AI ( 40+ preferred) all have published rules the firm meets. It may still sign with Mercor. It will know what the listing's range means for its own files.

Questions people ask

Does Mercor buy company data?

Yes. Mercor connects to a company's tools, extracts and anonymises its records with its own automated pipeline, at its own cost, and licenses the resulting dataset to AI labs, paying by wire typically 2 to 4 weeks after signing. Its data page names eight target industries and its referral listings price 25 kinds of company.

How much does Mercor pay for company data?

Its listings put what a company earns at $10,000 to $1,500,000, across 25 kinds of company: for example $50k to $1.2M for a law firm, $100k to $1.5M for a manufacturer and $50k to $500k for a software company. Where a company lands within its range depends on volume and depth, and is set after Mercor reads the records; its data page says the more tools you share, the higher your payout.

What tools does Mercor connect to?

Its data page shows Slack, Teams, Gmail, Outlook, Salesforce, GitHub, Drive, Notion and Confluence first, and says it supports 100+ tools in all. Its listings add industry systems, from Procore and Revit to Epic, Relativity and QuickBooks. You choose which to connect, and it says no data leaves your systems until you approve it.

How does Mercor anonymise company data?

With its own automated pipeline, which it pays for: it says no AI lab buyer or third-party vendor ever sees your raw data and its engineers do not manually review it. It publishes no de-identification standard and does not say where processing runs, so ask both, and ask whether an independent party checks the result. In the Spirit auction Mercor bid 35% more for the right to use its own tools rather than an independent scrub.

Does Mercor take companies outside the US?

Yes, by its listings' own words: 18 of them say "Anywhere", three prefer the US, the accounting listing requires it and the call-centre listing names its markets. Its terms ask users to confirm they are not in a country under US sanctions, and its payment terms name no currency or payout country.

What did Mercor bid for Spirit Airlines' data?

In the August 2026 auction it bid $5.2 million with independent de-identification or $7 million with its own tools, and later $10 million on in-house terms. The estate preferred independent scrubbing, Google won at $10M, and Mercor's bid stands as the alternate at $7.5M; the sale's approval hearing is now set for 17 November 2026.

How do Mercor referrals work?

Anyone with a Mercor Data partner account can refer a company from the listing it fits; the first valid referral counts, a company already working with Mercor does not, and each introducer may have up to five companies awaiting review at a time. Mercor then contacts the company itself, and the company signs its own agreement with Mercor and pays nothing for the introduction. The introducer's payout is set on each listing and paid after the company's first delivery.

What this means for you

Mercor publishes more about price than almost any buyer, and less about its licence than most. That makes it easy to see whether a company like yours is worth its while, and hard to see what you would be giving for the money. Prepare for the second before you answer the first.

1. Find your listing. Read the tables above for the kind of company you are, and check its size, place and required records against what you actually hold. The buyer match tool does the same across every buyer's published rules.
2. Get your own reading of value. A listing's range covers a whole kind of company. Value your data against published prices and sales to see where yours is likely to land.
3. Decide the scope before you connect anything. Which tools, which years, and which records inside them stay out. Use the readiness check and the pages for your systems under rights.
4. Ask the pipeline questions. Where processing runs, how long the raw copy is kept, to what standard it is de-identified, and the current SOC 2 report; our de-identification standard is the checklist.
5. Ask for the licence in writing. Term, exclusivity, sublicensing and what survives the end, plus which event triggers payment. Put the written offer through the offer check and decide where to stand with the negotiation page.
6. Hear from more than one buyer. Apply to Mercor directly, or be introduced to it and every other buyer whose rules take your data with one press after your valuation. Compare price, licence, de-identification and payment together, as the guide to every company-data buyer programme and the buyer board do.