Weeks to months, from first inquiry to money in the bank. Here is each stage, what you provide at each one, what slows it, and why a two-week promise should make you ask more questions, not fewer.
Last checked: 7 October 2026. Buyer statements are quoted as published on that date.
A sale of company data to an AI buyer usually takes weeks to months. The deal moves through an NDA, a review of what you have, a written agreement, the export, de-identification and an acceptance check before any payment is due. Practitioners cite 60 to 90 days to close.
That range is not a promise. A small, well-prepared company with clean records and a quick lawyer may move faster; a firm with client confidentiality questions, several systems and a board to consult may take longer. What you can control is how ready you are before the first conversation.
It helps to think of two clocks. The buyer’s clock covers its review, its legal team, its export tooling and its de-identification work; you cannot speed it up. Your clock covers inventory, internal approvals, client contract checks, admin access and your own lawyer. On many deals the seller’s clock is the longer one, simply because nobody started it before applying. Everything in the preparation plan further down is about starting your clock early, so that the buyer’s steps are the only ones left when an offer arrives.
The order is broadly the same across buyers. Names and paperwork differ.
You apply or reply to a buyer. It checks its published criteria, such as employee count, location, language and years of records, and asks what systems you use.
unclear answers about systems and history; waiting to hear from the owner.
Both sides sign a confidentiality agreement before details or samples are shared.
routing a simple NDA through a lawyer who is busy elsewhere.
You describe what exists and share representative samples, not the full dataset. The buyer reviews quality, volume and fit.
no inventory yet; samples that contain things that should have been excluded.
Price, scope of use, exclusivity, warranties, indemnities, deletion, audit rights and payment terms are negotiated and signed.
redlines on exclusivity and indemnity; internal sign-off; checking client contracts.
The agreed scope is pulled from your systems, often with the buyer’s tools or guidance.
admin access, API limits, exports that miss history or attachments.
Names, contact details and other identifiers are removed or replaced. Mode, for example, says it de-identifies before onward delivery; micro1 says sensitive and confidential information is scrubbed (both as published).
more personal data than expected; rounds of spot-checks.
The buyer checks the delivered data against the criteria in the agreement.
vague acceptance criteria; gaps found after export.
Paid as the agreement says: one-off, in milestones, or recurring.
milestones tied to steps you do not control; invoicing and payment terms.
Knowing which stages depend on you shows where preparation pays off.
| Stage | Mostly set by | Your lever |
|---|---|---|
| Inquiry and fit | Buyer | Clear, complete answers about systems, staff and years. |
| NDA | Both | A lawyer ready to read a short document quickly. |
| Manifest and review | You, then buyer | A manifest and clean samples ready before you apply. |
| Agreement | Both | Your limits decided in advance; signers available. |
| Export | You | Admin access arranged and a test export done. |
| De-identification | Buyer | Exclusions already applied, so less personal data reaches this stage. |
| Acceptance | Buyer | Written criteria agreed before signing. |
| Payment | Buyer | Payment terms and triggers that are clear in the agreement. |
A fictional company with one buyer and no client-confidentiality issues. Stages overlap: export can start while the last clauses are agreed, and de-identification can begin on the first batch. The longest bar is the agreement, which is normal and worth the time.
| Delay | Why it happens | What helps |
|---|---|---|
| Internal sign-off | Owners, partners, finance and sometimes a board each need to agree. | Name the signers before you apply. |
| Client contracts | Confidentiality terms with your own clients have to be checked before scope is final. | Review the biggest client agreements early. |
| Contract terms | Exclusivity, resale, indemnity caps and audit rights take real negotiation. | Read the agreement guide first and decide your limits. |
| Export access | Admin rights, API limits and missing history in standard exports. | Test an export of one team or project in advance. |
| Personal data surprises | Real records contain more names, numbers and side remarks than expected. | Spot-check samples yourself before the buyer does. |
| Acceptance disputes | Criteria that were never written down clearly. | Ask for written acceptance criteria before signing. |
| A second offer | Waiting for another buyer’s review adds weeks. | Usually worth it; see getting more than one offer. |
Posts promising a signed, paid data deal within two weeks skip steps that someone still has to do. Count them. Your own lawyer needs time to read an agreement that covers warranties, indemnities and deletion. The buyer needs time to look at samples and decide what the data is worth. A real export from several systems needs checking. De-identification of years of messages or tickets is not instant, and acceptance only starts after delivery. Each of those can take a week or more on its own, and several of them cannot run in parallel.
Speed also has a cost. A seller rushing to close is the seller least likely to push back on a perpetual exclusive license, an uncapped indemnity or vague acceptance criteria. Time-limited offers work the same way: a deadline that leaves no room for your lawyer is a reason to slow down. Our list of questions to ask a data buyer is built for that moment.
Some buyers do publish speed statements. Grepped, for example, publishes “get paid in 7 days” (as published, checked 7 October 2026). Read any buyer’s timing statement as that buyer’s own description of its process, and ask the same question every time: which step starts the clock, and which step stops it?
Court-supervised sales follow a different calendar again. The reported Spirit Airlines data sale ran through bankruptcy proceedings, with bids and court steps that an operating company does not face; see our Spirit Airlines case page.
Payment is the last stage on the timeline, but several obligations keep running afterwards, and they belong in your planning too:
Build an inventory of systems and years, draft an exclusions list, check retention policies, run a secrets scan, name your internal signers and line up a lawyer. Our guide to preparing your data for sale covers each step.
Do not send a full dataset before a price, skip your lawyer, accept unclear acceptance criteria or agree to exclusivity you have not priced. Those shortcuts save days and can cost far more later.
A suggested order, not a requirement. Smaller companies may finish sooner; firms with client confidentiality questions may need longer. None of it involves sending data anywhere.
List every system, its administrator and the years it covers. Check auto-delete settings in email, chat and meeting tools so you do not promise history you no longer hold. Name the people who must approve a sale.
Draft a plain-language scope and an exclusions list. Pull your largest client agreements and check their confidentiality terms. Brief a lawyer so the agreement review later starts from context, not from zero.
Run a secrets scan and rotate what it finds. Try an export of one team or project to see what the tools include and miss. Choose representative samples and read them for anything that should have been excluded.
Write the manifest. Decide your limits on exclusivity, indemnity and payment structure before anyone makes an offer. Draft the message to employees. Then apply to more than one program.
A 25-person company has fewer systems and fewer signers, but often no in-house lawyer, so the agreement review depends on an outside adviser’s calendar. A 500-person company has more people to consult and more systems to export from. Data type matters as much as size. Documents and procedures are usually quicker to scope and clean than email or chat. Code adds a full-history secrets scan. Recordings add consent questions. Anything touching clients under confidentiality terms, patient information or staff and customers in the EU adds a legal review that cannot be skipped.
Interest after an application is the start of review, not an offer. Plans built on an early “yes” tend to slip.
A full dataset sent early does not shorten the process much, and it gives away your strongest position before there is a price.
If an offer expires before your lawyer can read the agreement, ask for an extension. A buyer that will not allow time for review has told you something.
Starting the export early feels efficient. If terms then change, you have pulled data you may not be allowed to deliver.
Year-end closes, audits, holidays and board meeting dates all affect when signers are available. Plan around them.
An IT admin running a large export will notice. Employees who hear about the deal that way slow everything down with fair questions.
Quoted from each program’s own pages, checked 7 October 2026. None of this is a promise of acceptance, amount or timing.
micro1 lists 30+ employees, mature operations, documented processes, modern software tools and primarily English, with US companies prioritized. Mode lists 20+ full-time US office employees (accounting firms 10+, law firms 6+) and several years of records the company owns. Grepped lists any vertical. Applying to more than one at the same time is how a second offer arrives without adding months.
Independent site. Some links are referral links: if your company signs with a buyer through them, the buyer may pay us a fee. You are not charged, and we never see your data.