The questions to ask first.
Business records, not gig work or consumer personal data. Public evidence can guide scoping; your agreement controls the actual terms.
Does micro1's $1B commitment guarantee an offer for our company?
No. micro1 announced a commitment to acquire and license enterprise operational data over the next 12 months. That is a company spending announcement, not completed expenditure, an acceptance rate or an offer to your business. Its published enterprise FAQ still targets mature companies with 30+ employees and established documentation. Ask about your specific workflow, current demand and a scoped written offer before transferring records.
Are the Spirit Airlines data bids verified seller payouts?
No. The August 14, 2026 auction notice records Google's $10 million successful bid and Mercor's $7.5 million alternate bid, subject to court approval. micro1's September 3 notice proposes $12.5 million. Those filings document bids, not funds received. This research did not establish subsequent closing or payment, and a bankruptcy asset is not a small-business price comparable.
Why would an AI company pay for an ordinary business’s records?
Some programs seek records that show how real work is performed: procedures, approvals, escalations, decisions and outcomes. micro1 and Appen explicitly describe this demand. That creates a potential revenue opportunity from existing records, not a guarantee that your business qualifies or that a buyer wants every file you hold. Start by describing a useful workflow and the permissions you control.
Is this about doing paid AI training tasks as an individual?
No. These articles address authorized businesses licensing existing operating records. The company-data programs are distinct from expert hiring, interviews, contractor work and referral rewards on some of the same providers’ websites. Check the data-program page and proposed company agreement rather than relying on worker payment terms or hiring reviews.
Can a company with fewer than 30 employees participate?
Possibly in another program, but do not assume acceptance. micro1’s Enterprise Data Partnership targets 30+ employees. Handshake’s intake seeks 20+ full-time employees and at least three years in operation. A provider that publishes no minimum has not confirmed that it accepts your company. Ask about size, history, location and current demand before preparing exports.
Does meeting the employee minimum guarantee an offer?
No. micro1 couples its size target with mature operations, established documentation and high-quality records. Handshake says any payout depends on underlying sources, volume, quality, due diligence and acceptance of licensing terms. Treat headcount as an early program-fit check, not a valuation formula or promise of participation.
Can a business outside the United States apply?
Eligibility is program-specific. micro1 says it prioritizes US companies followed by other Western markets, and primarily seeks English operations and documentation with English-speaking representatives. Global referral acceptance is not global company acceptance. Ask the relevant data team about your seller country, record languages, processing locations and current demand; do not infer approval from a country menu.
Can a winding-down business license records?
Some programs explicitly discuss winding-down companies. Appen includes them in its enterprise-data process. Centific says winding-down companies may use a dedicated asset-purchase agreement. Confirm authority, retained access and the proposed transaction type before systems are shut down; a wind-down offer may be a sale rather than an ownership-retaining license.
Which kinds of records might be useful?
Published programs describe SOPs, knowledge bases, support and IT tickets, CRM histories, project records, messages, code and finance workflows. Usefulness depends on what they show, not just their format. Linked records that explain an approval or resolved exception may be more relevant than an isolated final document. Every category still needs ownership, confidentiality and privacy review.
Do we need to provide source code?
Do not assume so. micro1 lists procedures, documentation, CRM and project histories. Appen lists several non-code operational categories. Turing’s Project Lazarus explicitly says code is not required and is not bought on its own. Ask the prospective partner which records support the current use case; exclude code if you cannot clear its IP, licenses or secrets.
Should we upload a sample in the first inquiry?
Our recommendation is to begin with metadata: workflow description, system names, period, approximate counts and unresolved restrictions. Do not send raw records before establishing authority, recipients, purpose, handling and retention. An NDA should not be treated as permission to disclose client-owned or otherwise restricted material. Have an authorized reviewer approve any later sample.
Should we clean out failed tasks and reopened tickets?
Not simply to make the archive look better. Appen explicitly identifies failure and recovery paths, rejected approvals and reopened tickets as useful signals. Document missing or inconsistent records honestly and distinguish original history from later annotations. Preserve the source history under existing controls; transformation should follow an agreed scope rather than create a misleading success-only dataset.
What should a first inventory include?
Use a workflow, system, date range, approximate count, link quality, known gaps and responsible internal owner for each candidate collection. Maintain a separate exclusions and rights-review register. Do not embed confidential examples, customer identities or credentials in the inventory. This is a preparation proposal, not evidence that the records are legally cleared or that any buyer will accept them.
Are exports safer than connecting our systems?
Not inherently. Troveo describes guided exports without OAuth; Mercor describes scoped, read-only connections. Compare the actual boundary, raw-data recipients, processing locations, logging and sample approvals. A read-only connection can still copy records, while an export can accidentally contain sensitive attachments or history. Choose the method only after its specific controls and responsibilities are understood.
Can LicenseTheData.com estimate our expected income?
No reliable expected-income figure is supported by this research. Public tiers and calculators are provider marketing or indicative estimates, not achieved payout distributions for ordinary businesses. Handshake expressly says its estimate is not an offer. Evaluate workflow depth, demand, rights and preparation effort, then request a scoped written offer instead of relying on a headcount or storage-size multiplier.
What affects a buyer’s valuation?
Published programs describe factors such as rarity, structure, history, size, quality and uniqueness. Appen and Centific emphasize connected records and operational context. Those are useful dimensions to describe, but they do not supply a universal dollar schedule. Ask how the particular buyer values the scope it is requesting and whether exclusions change its offer.
Are there setup or preparation fees?
Appen, Mercor and Centific state that they cover extraction, anonymization and transfer without setup fees or deductions. Confirm those claims in your written offer. Do not extend them to every provider or assume that internal staff time and outside advice cost nothing. Ask for the complete gross-to-net calculation, including any commission, recurring share or bank-charge allocation.
Does Mercor pay 2–4 weeks after we contact it?
No. Its published FAQ anchors the typical 2–4-week interval to a signed agreement and says wire payment is issued upon delivery of the anonymized dataset to the buyer, depending on extraction and processing time. Initial qualification and negotiation occur before that anchor. Confirm the applicable payment obligation, delivery conditions and due date in your agreement.
Does Appen’s two-business-day statement mean a quick payout?
No. Appen describes a response within two business days. It separately says terms are agreed before extraction and payment is on acceptance rather than submission. Ask who accepts the contribution, what criteria apply and when payment becomes due after acceptance. Do not use the response interval as a cash-receipt forecast.
Does Centific’s 6–10-week interval guarantee cash by week ten?
No. Centific says most partnerships reach a first accepted contribution within six to ten weeks and describes payment on acceptance. That is not a verified bank-receipt deadline for every seller. Ask for the acceptance definition, contractual payment due date and any invoicing requirements; track accepted contribution and received payment as separate events.
Is licensing revenue recurring?
Only if the actual structure provides for it, and later payments may depend on later sales. Turing describes a single payout for Project Lazarus; Replay advertises upfront cash plus revenue share; Troveo describes earnings on repeat sales. Compare definitions, reporting, deductions and due dates. Do not treat possible repeat transactions as guaranteed monthly income.
What percentage does Troveo pay the business?
A currently verified revenue-share percentage was not established in this research. Troveo publicly says licensors share in each sale, but that is not enough to calculate net proceeds. Ask for the current agreement and any rider, including the revenue definition, share, deductions, reporting and payment deadline. We do not use indexed contract clauses from a page that returned a 404 as verified terms.
Do we retain ownership when we license records?
micro1 and Troveo publicly say underlying ownership remains with the company. Confirm the exact grant in your agreement. Retaining ownership does not itself establish non-exclusivity, narrow rights, cancellation rights or limited downstream use. Read the provisions for prepared data, derivatives, sublicensing, term and rights that survive termination.
Can we keep using the records in our own business?
Ask for express confirmation in the proposed agreement, including continued operations, product development and future licensing. Do not infer unrestricted reuse solely from an ownership-retention statement. Check whether exclusivity attaches to source records, a prepared version, future updates or particular uses. If internal rights are essential, have them recorded as explicit carve-outs.
Is anonymized data automatically safe to license?
No. Determine what transformation actually occurs, whether identification risks remain and whether IP or confidentiality restrictions still apply. The UK ICO says pseudonymised personal data remains within data-protection law and is not the same as anonymised data; its guidance is under review. That UK guidance does not settle every jurisdiction’s requirements. Get an assessment for the specific dataset and use.
Do employees or customers need to consent?
There is no universal answer supplied here. Have qualified advisers assess applicable law, notices, lawful bases, contractual restrictions and the proposed new purpose. Do not assume that collecting information during ordinary operations authorizes AI licensing, or that consent is always the only available basis. Pause transfer until the relevant obligations and permissions are resolved.
What if the records include client work or confidential material?
Separate those records for review before sending a sample. Identify ownership, customer contract restrictions, professional obligations and any permission needed for the particular use. A confidentiality agreement with a buyer does not automatically override your obligations to a client. If those restrictions cannot be cleared, exclude the material or decline the transaction.
Will an AI lab see our raw records?
Mercor, Appen and Centific publicly say labs or third parties do not see raw data in their described processes. Those are provider claims, not security audits performed by LicenseTheData.com. Ask who operates extraction and transformation, whether anyone can review content, and what your actual agreement permits. Other workflows should not inherit those assurances by association.
Can we cancel the license and undo model training?
Do not assume either right. Ask separately about terminating future use, deleting raw and prepared copies, onward recipients and treatment of derivatives or completed training. An ownership-retention statement does not promise revocability. Seek explicit contractual provisions; do not promise that deleting source records reverses uses that have already occurred.
What should we ask about exclusivity?
Define its records, uses, buyers, territory, duration and treatment of future updates. Ask what the partner commits in exchange and whether there is an exit if a sale does not happen. Consider internal reuse and future buyers. These are negotiating questions, not protections we have confirmed in every provider’s agreement.
Is a marketplace listing the same as a buyer offer?
No. Datarade offers provider discovery and listings. Nomad describes routing buyer requests and direct negotiation. Neither a listing nor a matched lead establishes a completed deal or an obligation to buy your records. Ask who scopes, prepares, contracts for and pays for the dataset, rather than treating every route as a managed acquisition program.
How should we compare two written offers?
First align the records and rights each offer purchases. Then compare guaranteed and conditional cash, deductions, acceptance criteria, repeat-sale accounting, exclusivity, permitted uses, retention and liability. Keep staff and adviser effort separate. A larger gross amount is not necessarily better if it requires broader rights or unresolved handling obligations. This is an editorial comparison framework, not financial or legal advice.
When should a business decline to license its records?
Decline or pause when authority is uncertain, client restrictions cannot be cleared, privacy risks cannot be acceptably addressed, required exclusions destroy useful context, or the proposed rights would damage the business. Also stop if the effort exceeds your budget or no acceptable offer emerges. Additional revenue is a reason to investigate, not a reason to bypass obligations.
Are these provider claims verified seller reviews?
No. These articles distinguish primary program descriptions from received-payment evidence. The underlying research did not locate independently evidenced named small-business operating-data payouts for the core programs reviewed. That is a research limitation, not proof that nobody was paid. We do not repurpose hiring reviews, provider-hosted testimonials or aggregate licensor totals as independent SMB payment proof.
What does the October 13, 2026 publication date mean?
It is the publication date assigned to LicenseTheData.com’s educational draft articles. It does not imply that a provider launched, a payment occurred or a contract was legally reviewed on that date. Eligibility and commercial terms can change. Recheck the linked primary pages and obtain current written terms before making a decision.