Services
Private AI Deployment
AI on your hardware. Your data never leaves your infrastructure.
Most AI deployments require your data to pass through infrastructure you do not control. For organisations handling sensitive data (patient records, legal documents, financial data, confidential client information) that is not an acceptable trade-off. A private deployment runs entirely within your environment. The models run locally. The data stays put.
What you get
- Working deployment — A functional AI system, tested and running on your infrastructure before handover. Not a proof-of-concept. A production-ready deployment with the rough edges resolved.
- Model selection and configuration — We select and configure open-weight models appropriate for your use case. You are not locked to a single provider. Models can be changed or updated without renegotiating a contract.
- Runbook — Documented procedures for day-to-day operations: how to start and stop the system, how to update models, how to respond to common failure modes, how to monitor health. Written for the people who will actually run it.
- Admin documentation — Configuration reference, integration map, environment variables, access control setup, backup and recovery procedures.
- Handover session — A structured session with your technical team to walk through the deployment, answer questions, and confirm they can operate it independently. We do not leave until you can run it without us.
- 30-day post-handover support — Access to us for questions in the month after handover. Not a managed service. A safety net.
Process
- 1
Scoping call (30 minutes)
We establish the use case, the data types involved, the infrastructure you have available, and the regulatory constraints that apply.
- 2
Scope agreement
Fixed deliverables, fixed price, defined timeline. Signed before work begins.
- 3
Deployment and configuration
We build the deployment in your environment. Access is by mutual agreement. We do not retain persistent access after handover.
- 4
Testing
We test against the use cases agreed in scope. You test against your own acceptance criteria.
- 5
Documentation and handover session
You leave with everything you need to operate and maintain it.
When it is the right fit
You should book a scoping call if:
- Your organisation handles data that cannot leave your jurisdiction: NHS patient data, legally privileged documents, financial records subject to data residency requirements.
- You are currently using a cloud-hosted AI service and you have received questions from your DPO, legal team, or a client about where the data is processed.
- You want to add AI capability to internal workflows (document analysis, summarisation, Q&A over your data) without sending that data to a third party.
- You have tried a cloud AI service and the results were acceptable, but the data handling was not.
Data residency, model choice, and jurisdiction
Data residency is customer-controlled. The deployment runs on your infrastructure, in your chosen jurisdiction. If you are a UK organisation and your data must stay in the UK, it stays in the UK. If you operate in the EU and need to comply with EU GDPR data transfer restrictions, the deployment is scoped accordingly.
Model selection is yours. We work with a range of open-weight models suited to different tasks: text analysis, summarisation, structured extraction, Q&A. We recommend on the basis of your use case and your hardware. You are not required to use any specific provider or pay any per-token fee after deployment.
Book a scoping call
30 minutes. No obligation. We will tell you if it is not a fit.
Book a scoping call