How much is your company’s operational data worth to AI?
This guide is for owners and managing directors who have heard that AI companies pay for real business know-how and want a sense of scale. The short version: one programme publishes a range of$100K to $2M+ for approved company data packages, and where a company lands depends on what it holds, which no published figure can see.
How much is operational data worth to AI?
There is no market price. One programme publishes a range of $100K to $2M+ for approved company data packages (checked 6 Oct 2026). Where a company sits in that range depends on what it holds and how clean it is. Any single figure is an estimate, not an offer.
Operational data here means the working know-how of a business: procedures, playbooks, ticket and project histories, quality records and similar. Our guide to what data AI labs buy sets out the types, and how AI data programmes work explains who decides what is paid.
What do published and reported figures say?
Very little is public. The clearest published number is a programme’s own range for approved company data packages. Press reports of smaller sums concern archives from closed companies, which is a different thing from the live operating knowledge of a working business, so treat them as context, not a benchmark.
On 18 April 2026 Slashdot quoted a Fast Company report on SimpleClosure, a service that helps companies shut down. It said the firm had processed about 100 deals in the past year and that payouts ranged from $10,000 to $100,000 (Slashdot, 18 April 2026; we could not open the Fast Company article itself). On 3 September 2026 SimpleClosure’s own page described deals “in the tens to hundreds of thousands of dollars for unique assets” (SimpleClosure, 3 September 2026). Those are wind-down sales of old workspace archives. They tell you the category exists, not what a live company would receive.
What does the estimate table show?
It shows an illustrative range for each team size and number of years operating, the same figures as our interactive estimate. It is an estimate, not an offer, a valuation, a price or a promise of payment, and it moves only with those two inputs.
| People \ Years operating | Under 3 | 3–7 | 8–15 | 15+ |
|---|---|---|---|---|
| 20 people | $110K–$170K | $170K–$220K | $220K–$280K | $280K–$330K |
| 100 people | $220K–$330K | $330K–$440K | $440K–$560K | $560K–$670K |
| 250 people | $330K–$500K | $500K–$670K | $670K–$830K | $830K–$1M |
| 500 people | $440K–$670K | $670K–$890K | $890K–$1.1M | $1.1M–$1.3M |
| 1,000+ people | $670K–$1M | $1M–$1.3M | $1.3M–$1.7M | $1.7M–$2M+ |
Estimate only: not an offer, a valuation or a promise of payment. Based on team size and age, and scaled to one AI data programme’s published range for approved company data packages ($100K–$2M+, checked 6 Oct 2026). Each programme decides whether to work with a company and what it pays. Some programmes set a minimum team size.
The table cannot see your records. For a reviewed answer based on what you actually hold, use the three-minute check.
What drives the price up or down?
Six things matter most: how unique the knowledge is, how deep the history runs, how complete and consistent the records are, how much is repeatable workflow rather than one-offs, how cleanly personal and client information can be removed, and scope and exclusivity. We do not know the weights any programme uses.
- Uniqueness. Specialist judgement that is hard to find elsewhere is worth more than widely available practice.
- Depth of history. Years of records show how work changes over time, which is why age is an input above.
- Completeness and consistency. Records that follow the same structure and cover the whole process are easier to use than scattered notes.
- Repeatable workflow. Many examples of the same task done well beat a pile of one-off projects.
- Clean removal of personal and client data. If names and client details are woven through everything, the usable part shrinks. Read client consent and licensing before you start sorting.
- Scope and exclusivity. What you license, for how long, and whether you grant it to one buyer only all change the terms.
Preparation helps with several of these. Our guide to licensing your SOPs to AI covers what to gather first.
Why are labs paying a premium now?
Real company know-how is scarce, and labs want it to train and test AI that does workplace tasks. As more companies come forward, we expect prices to settle, so the strongest terms are likely to go to those who move early. That is our expectation, not a deadline, and we can’t say exactly when.
Nothing here is a reason to rush a decision. Agree terms you understand, and keep your own legal advice in the loop. This is general information, not legal advice; a solicitor can advise on your contracts and your obligations under UK GDPR (see our privacy notice for how we handle your own details).
How do you get a figure for your own company?
Take the three-minute check. We review your answers ourselves and reply with what fits, if anything, including what we know of each programme’s published terms. The check is free; we charge a success fee only if a deal closes, agreed in writing first.
We act for your company, not for a buyer. Programmes may also pay us a referral fee, which we tell you about first and which does not decide what we recommend. The full wording is on our disclosures page. You can also browse all our guides.
Common questions
Is the table an offer?
No. It is an illustrative estimate, not an offer, a valuation, a price or a promise of payment. It is scaled to sit inside one programme’s published range, and each programme decides whether to work with a company and what it pays.
Does a bigger team always mean a higher figure?
In the estimate, yes, because team size and years operating are the two inputs. In practice we expect the records to matter as much as headcount: a small team with deep, consistent, well-documented work may interest a programme more than a larger one with thin or scattered records.
Will prices stay where they are?
We can’t say. We expect prices to settle as more companies come forward, so the strongest terms are likely to go to those who move early. That is an expectation, not a deadline, and we can’t say exactly when.
What does Orca charge?
The check is free. If you go ahead, we act for your company and charge a success fee only if a deal closes, agreed in writing before anything starts. Programmes may also pay us a referral fee, which we tell you about first. Neither decides what we recommend.