What owners ask before they say yes.

Operational records: the email, chat threads, tickets, CRM updates, project notes and documents that show how your team gets work done. AI labs want to see the whole path of a piece of work, from the first client email through the internal discussion, the options you rejected, the fix, and the reply.

It is not your customer list, your pricing or your trade secrets. Sensitive categories such as privileged legal material, patient health information, payment details, government ID numbers and minors' data are excluded.

No. Deals in this market are licenses. You keep ownership of your records and keep running your business exactly as before. The buyer's agreement spells out what is licensed, for how long, and on what terms, and your counsel reviews it before anything is signed.

Octet never asks for files, exports, passwords or logins, and never handles your data. Our questions are about which systems you use, not what is inside them.

If a buyer goes ahead, your own team typically runs a one-time historical export under the buyer's agreement. There is no ongoing access to your systems.

Reputable buyers have records de-identified by specialist third parties before their researchers work with them. Names, emails, phone numbers and addresses are replaced with consistent placeholders ("Employee 632", "Company A"), so the records still show how work moved without identifying anyone.

Don't try to de-identify the data yourself first. Done by hand it usually destroys the relationships that make the records valuable.

It varies widely. The biggest drivers are how many people work in your tools every day, how many years of history sit in the systems you use today, how much of your work is written down rather than done by phone, and how many systems hold it.

Our estimate gives an illustrative range based on public market information. It is not an offer. Only a buyer can price a dataset, after its own review, and many companies are not accepted.

Data buyers pay us a referral fee if a deal we introduced closes. You never pay Octet anything, and there is no cost to get an estimate or talk to us.

We work with several AI labs and data buyers. Buyer identities are confidential, so we don't list them publicly. Before any introduction we tell you which buyer we recommend and why, and nothing about your company is shared without your go-ahead.

Buyers commonly ask for a period of exclusivity on the dataset they license. Be wary of perpetual terms. If you might sell the business in the next few years, it's worth timing any exclusivity to end before then, so a future owner inherits an asset rather than a restriction.

A data license is something an acquirer will ask about in due diligence, so keep the agreement on file. A license payment is typically one-time, so don't expect it to be treated as recurring earnings in a valuation.

The estimate takes a few minutes. After that it's usually a short fit call and, if a buyer proceeds, a review and a one-time export run by someone on your team. Timelines are set by each buyer and vary.

Responsibilities and protections are set out in the buyer's data agreement. Read it with your attorney before signing, and ask the buyer how de-identification is done and verified.

Buyers' requirements move over time. If you've agreed to hear from us, we'll get in touch if your company becomes a fit. You can also come back and run the estimate again whenever your systems or team change.