Frontier labs are paying for real operational knowledge, and the European half of that market is largely unserved because the compliance work is hard. We do that work, so your company can licence anonymised operational data without leaving EU jurisdiction.
Models are short of one thing above all: records of how competent organisations actually do work. Not published articles, not scraped text. The internal record of decisions, exceptions and judgement that only exists inside operating companies.
Standard operating procedures, playbooks, internal knowledge bases, training material, quality manuals.
Ticket and case histories, CRM and workflow metadata, project histories, review and approval trails, the reasoning behind outcomes.
Your specialists evaluating model output in their own domain, which is the highest value contribution and the one only your people can make.
Most European companies could licence operational data tomorrow and do not, because the legal path is genuinely unclear and the downside sits with them. Establishing lawful basis, handling any personal data, satisfying employee representation, keeping processing inside EU jurisdiction, and documenting all of it well enough to survive scrutiny is the actual work. That is the part we own.
We assess your data against GDPR, record the basis for processing, and document it in a form your DPO can sign off rather than argue with.
Data is handled by EU entities under EU law, with residency and access controls written into the agreement, not asserted in marketing copy.
Article 10 names annotation and labelling as regulated operations for high risk systems. The provenance and governance record we produce is built to that standard, and it is yours to keep.
Where works councils or employee consultation apply, that process is part of the programme and runs before contribution starts, not after someone complains.
Nothing enters the programme without your explicit sign off on that category of data. The default is exclusion.
Identifiers are stripped and records are reviewed before anything is used. Where structure matters more than content, records can be rewritten to preserve the workflow shape.
You licence a defined use. You do not sell your data, and you do not lose the right to use it yourself in any way.
You can name parties your data may not reach, and you can end future contribution on notice without unwinding what is already delivered.
Scroll through the process. Nothing commits you until the licence itself.
A short call and a look at what you hold. We come back with what is licensable, what is excluded and why, and an indicative value. No cost and no commitment.
Lawful basis, scope, exclusions, residency and any employee consultation are settled and written down. Nothing moves until this is signed off.
We do the work. Anonymisation, structuring and quality control run on our side, designed to take as little of your team's time as the data allows.
The dataset is licensed under the agreed terms and you are paid. Recurring contribution continues under the same agreement if you want it to.
Value comes from how much usable operational record you hold, how deep it goes, how unusual the domain is, and whether you licence exclusively. This gives you an order of magnitude in about a minute.
This is a model, not an offer. It reflects typical dataset scale for a company of this profile and the licensing structure you chose. What you would actually be paid is set by the assessment, in writing, before you commit anything.
Below roughly thirty, how unusual the domain is matters more than how much there is, so specialist teams are still worth a conversation.
Procedures, knowledge bases or case history that a new joiner could actually learn from.
This programme is built for EU and EEA companies, and for UK companies working to equivalent standards.
Volume helps, but it is not the deciding factor. A small team in a narrow regulated field can hold more valuable operational knowledge than a large generalist one.
No. Personal data is removed before use, and categories that cannot be safely anonymised are excluded from scope at the assessment rather than handled later. If your data cannot be made lawful to licence, we will tell you that and stop.
It depends on the data, which is why the assessment comes first. We establish and document the basis for each category in scope, and your DPO reviews it before anything moves. Where no basis exists, that category does not enter the programme.
You keep ownership. You are granting a licence for a defined use, with named exclusions if you want them, and you continue to use your data exactly as you do today.
You can name parties your data may not reach, and that exclusion is written into the agreement. Exclusive licensing is also available and is priced differently.
The extraction and preparation work sits with us. The time you spend is concentrated in the assessment and the governance step. Where your experts review model output, that is paid separately and scheduled around them.
Where employee consultation applies, it runs as part of the governance step before contribution starts. We would rather lose weeks there than have a programme unwound later.
Yes. You can end future contribution on notice. Datasets already licensed and delivered continue under their agreed terms, which is why the scope conversation happens up front.
A short call and a look at what exists. You get a written view of what is licensable, what is excluded and why, and an indicative value. Nothing is committed and nothing moves until you have that in hand.
Choose a slot below and we will come to the call having already looked at what you described. If nothing suits, we will reply within one working day.
The scheduling calendar is provided by Calendly and only loads once you choose to book, so nothing third party runs on this page before then.