If you run an agency, a consultancy or a dev shop, your company's most valuable data was mostly produced for someone else. The proposals, the briefs, the code, the decks and the deliverables all sit in your systems, and a lot of them belong to your clients. That doesn't rule you out of licensing your data to AI labs. It changes what's in scope. Read on to find out what a services firm usually owns, what its clients usually own, what buyers pay for in this kind of data, and how to license your firm's operating history without touching a single client deliverable.
Why AI labs want to see how a services firm works
AI labs are training models to do work, and services firms are where a lot of work gets done. A model that will scope a project, write a brief, revise a proposal after feedback, or turn a spec into working code needs to have seen thousands of real examples with real outcomes. The public web has templates and advice. It doesn't have your 300 proposals with the ones that won marked, your project channels where the scope changed twice, or the postmortem that explains why a delivery slipped.
Buyers are paying six figures for full-company datasets that show real work in a real business, and a services firm with 50 or more people and a few years of history holds that kind of record. Our guide to enterprise operational data explains what buyers want from it, and our page on business data covers the categories we license.
What you own and what your clients own
Go system by system. For each one, the question is whether the record shows how your firm does its work, or whether it's something a client paid you to make.
| What's in your systems | Who usually owns it | Can you license it? | What has to come out |
|---|---|---|---|
| Deliverables made for a client (decks, reports, creative, designs) | The client, under most contracts | Usually no | Everything, unless the contract says otherwise |
| Code written for a client as work for hire | The client | Usually no | Everything, unless the contract says otherwise |
| Your own frameworks, templates, internal tools and reusable libraries | You | Yes | Client names, credentials |
| Proposals, quotes, statements of work, pitches | You | Yes | Client names, contacts, client-specific pricing |
| Project records (tickets, timesheets, Slack channels, project tool history) | You | Yes | Client identities and personal details, replaced with pseudonyms |
| Internal playbooks, SOPs, training material, postmortems | You | Yes | Names, anything specific to a named client |
| CRM and sales pipeline history | You | Yes | Contact details and client identities |
| Material and data your clients gave you | The client | No | All of it |
The first two rows are the ones most firms worry about, and the answer is usually that they stay out. The rest of the table is the record of how your firm runs, and that's what buyers pay for. A proposal shows how you scope and price work. A project channel shows how a team handles a change of scope. A postmortem shows what went wrong and what changed. None of that is a client deliverable, even when a client's project is the subject. Our guide to what your company's docs are worth to AI labs goes folder by folder through the documents side of this.
This isn't legal advice. Your contracts decide what you own, and a lawyer should read them before you license anything.
The three contract clauses to check
Most client agreements answer the ownership question in three places, and you'll want to read all three before you scope a license:
- The IP assignment or work-for-hire clause. It usually gives the client the deliverables and the code written for them, and usually leaves your pre-existing IP, tools and methods with you.
- The confidentiality clause. It usually covers the client's business information and anything they gave you, which is why client-provided material stays out and client identities are replaced with pseudonyms.
- The data or records clause, if there is one. Some contracts say who owns the project records themselves. Most don't, and the records of how your team worked stay with you.
If a contract is silent or unclear, leave that client's project out of the first scope and come back to it later. Nothing about this process starts with "license everything."
What consultancy code is worth
Dev shops and technical consultancies ask about code first, so here's the plain answer. Code written for a client under a work-for-hire clause belongs to the client and stays out. Code you own (internal tools, reusable libraries, frameworks you've built across projects, products you've spun out) can be licensed, and buyers pay for it the same way they pay for any codebase, on the strength of its history. The commits, the reviews, the issues and the fixes are the asset, and a library you've maintained across dozens of projects has a longer and richer history than most startup repos.
Code on its own sits at the small end of the market. Public closure-market deals have run from about 5,000 dollars for a single repository to 10,000 to 100,000 dollars for a startup's whole archive. The six-figure deals are for the full record of an operating company, and that's the scope a services firm should aim at, with the code as one system among several. Our guide to selling source code to AI labs covers how buyers price code and what to export.
Determining your data scope
Start with your own operations and leave client deliverables out entirely. For most firms, the first scope is:
- Proposals, quotes and statements of work, with client names removed
- Project tool history (Jira, Asana, Monday or similar), with client identities replaced by pseudonyms
- Internal chat channels about projects, staffing and process
- Playbooks, SOPs, templates, training material and postmortems
- CRM and pipeline history, with contact details removed
- Your own code, tools and libraries
Names, contact details, credentials and client identities are replaced with pseudonyms to a documented standard before a buyer sees anything, and that's our work, done during packaging. Anything a client gave you, anything under a confidentiality agreement, and anything in an HR, payroll or personal folder stays out. Our owner's guide to licensing company data for AI walks through the exclusions in detail.
What makes a services firm's data worth more
Six things decide what your data earns, and volume by itself does little. A 40-person consultancy with ten years of specialized records can be worth more than a big company with generic data. The six, in order:
- How unique it is. A firm that does one thing well in one industry holds records nobody else has.
- How many years of history you hold.
- How connected your systems are. A proposal linked to the deal in your CRM, the project channel and the postmortem is worth more than any of them alone.
- Whether your records show outcomes. Which pitches won, which projects shipped on time, which didn't.
- How clean the rights are. Clear contracts and a clean split between your work and client work move fast.
- What buyers want right now. Today that's the record of how professional work gets scoped, revised and delivered.
Our guide to how much your company's data is worth to AI labs goes through each factor with the price points behind it.
How the export and the license work
The process runs in four steps. Scope, where you take the data value assessment and we agree which systems are in and which clients' work is out. Agree, where you sign standard terms that cover the license and your earnings. Package, where you export the data (a few hours, nothing installed) and we replace names and client identities with pseudonyms and document what's included. Get paid, where the first sale lands and you share in every sale after it.
You keep ownership the whole way through, your firm is never named publicly, and buyers are introduced to you only after an agreement is signed. The same dataset can be licensed more than once, so the first sale is the floor, and our guide to exclusive vs non-exclusive data licenses explains the choice you'll make on that.
Where to start
The quickest way to find out whether your firm's data is worth licensing is to take our free data value assessment (ten questions about your systems, history and industry, about five minutes). Answer for your own operations, leave client work out of your head while you do it, and you'll get an estimate of what the data could license for. A scoping call turns that into a real number.
Where Troveo fits
Troveo helps companies understand what data they hold, protect what matters, and license what's valuable, selectively. We've paid rights holders more than 20 million dollars across video, audio, gaming, robotics and business data, we work with more than 40 active buyers, and we take no fees and no deductions from your payouts. For a services firm, we scope around your client contracts, replace names and client identities with pseudonyms, package the data, and handle every buyer conversation, and you decide what's in and what's out. Start with the data value assessment, or talk to us about what your firm holds.