Communications
Threaded Teams, Slack and email conversations, with roles and timestamps preserved.
For AI labs
Rights-cleared UK operational data
English-language workflow data from UK companies, with documented provenance and de-identification. Manifest and redacted sample before you commit.
What UK company data can AI labs license?
Multi-year operational records from UK companies with 30 to 500 staff: threaded communications, documents and procedures, and system records such as CRM history, support tickets and finance workflows. The data is English-language, rights-cleared and de-identified, and suited to training and evaluating agents on real business workflows.
Documents and procedures
SOPs, tenders, method statements, contracts and reports, with version history where available.
Systems and records
CRM history, support tickets, project tasks and finance workflows, linked across systems where possible.
Accounting and financeLegalConstruction and engineeringSoftwareCustomer support and salesOperations and logistics
How do you know what you’re licensing?
Every dataset comes with a manifest covering sources, volumes, date ranges, de-identification and the lawful basis. You review the manifest and a redacted sample under NDA before you sign, so your legal and data teams can judge fit and provenance before any commitment.
Dataset manifestIllustrative example
| Dataset | SD-0142 |
|---|---|
| Sector | Construction and engineering |
| Headcount band | 101–250 |
| Date range | 2018–2025 |
| Sources | Teams, Outlook, SharePoint, CRM |
| Volume | 2.4m messages · 86,000 documents |
| De-identification | Consistent pseudonymous tokens |
| Lawful basis | Legitimate interests (LIA, DPIA) |
| Licence | Non-exclusive |
What documentation comes with each dataset?
The documentation your legal and data teams will ask for, prepared in advance: a manifest and redacted sample, a provenance record, the seller’s rights confirmation, a de-identification certificate and a UK GDPR summary covering the LIA, DPIA and transfer mechanism. See our compliance process.
Manifest and redacted sample
Sources, volumes, schemas and a representative sample.
Provenance record
Which company, which systems, extraction dates and chain of custody.
Rights confirmation
The seller’s confirmation that it can license each source.
De-identification certificate
The method used and the results of quality checks.
UK GDPR summary
Summary of the LIA, DPIA and the transfer mechanism used.
What are the licence terms?
Standard terms that work for both sides: non-exclusive by default, use limited to training and evaluation, no re-identification or resale, a fixed term with deletion or return at the end, English law, and pricing per dataset in GBP or USD. Custom scopes and time-limited exclusivity are available by agreement.
- Exclusivity
- Non-exclusive by default. Time-limited exclusivity by agreement.
- Permitted use
- Training and evaluation of AI models and agents.
- Restrictions
- No re-identification, no resale and no onward transfer of raw data.
- Term and deletion
- Fixed term agreed per licence, with deletion or return at the end.
- Transfer mechanism
- UK–US data bridge or the UK IDTA.
- Governing law
- England and Wales.
- Pricing
- Per dataset, quoted in GBP or USD.
How is the data delivered?
By encrypted transfer to your cloud storage after signing, in JSONL or Parquet, with threads, attachments and cross-system links preserved. Where a company takes part over time, refreshed data can be licensed on a schedule, so you can build longer-running datasets from the same source.
Formats
JSONL or Parquet, with threads, attachments and cross-system links preserved.
Secure transfer
Encrypted delivery to your cloud storage after signing.
Ongoing feeds
Some companies take part over time, so refreshed data can be licensed on a schedule.
How do AI labs work with Sterling Data?
In four steps: brief us on the domains, workflows, volumes and formats you need; review matching manifests and redacted samples under NDA; agree terms, which the seller approves before signing; then receive encrypted delivery with the full documentation pack.
- 1
Brief us
Tell us the domains, workflows, volumes and formats you need.
- 2
Review
We share matching manifests and redacted samples under NDA.
- 3
License
Agree terms. The seller approves the licence before it’s signed.
- 4
Receive
Encrypted delivery with the full documentation pack.
Questions from AI labs
Is the data personal data?
Identifiers are removed or replaced before delivery, and every licence bans re-identification. Where any personal data remains, it is processed on a documented lawful basis under UK GDPR, and transfers to US buyers rely on the UK–US data bridge or the UK International Data Transfer Agreement.
Do you sell the data or license it?
We license it. The UK company keeps ownership, and licences are non-exclusive by default, with time-limited exclusivity available by agreement. See selling vs licensing company data.
Which sectors are available?
Accounting and finance, legal, construction and engineering, software, customer support and sales, and operations and logistics. If you need something specific, we can source to order from UK companies that match your brief.
What do you need from us to start?
A short brief: the domains and workflows you need, approximate volumes, formats and intended use. We then share matching manifests and redacted samples under NDA.
Tell us what you need
Tell us the domains, workflows and volumes you’re looking for, and we’ll come back with matching datasets and what we can source to order.
Data partnerships: info@sterlingdata.ai.
Data you can defend
Rights-cleared, de-identified and documented under UK GDPR.