AI policy

01Our approach

FederatedOne exists to make private AI accountable. We bring together the agent platform, custom tools, GPU compute, the data centre and a trusted partner network, with one integrator answerable for the whole. This policy sets out the principles we apply to the AI systems we build, deliver and use.

02Scope

This policy applies to the four FederatedOne layers (FederatedOne Build & EcoPartners, FederatedOne Agentis, FederatedOne Compute and FederatedOne DataCenter) and to how we use AI in our own work. Specific commitments to a client are set out in our agreement with that client.

03Private by design

Our components are designed to keep AI and data under your control:

• FederatedOne Agentis runs on-premises, inside your own network.
• FederatedOne Compute uses enterprise NVIDIA GPUs in your own data centre.
• FederatedOne DataCenter runs your private cloud or on-premises build: hosting, power, cooling, networking and security, under your control.

We do not use client data to train models for other clients.

04Human oversight

People stay in control. Agents and routines in FederatedOne Agentis, and the tools we build with FederatedOne Build, operate under human approval gates: significant actions wait for a person to approve them.

05Transparency

We aim to make AI output understandable and checkable. Research in FederatedOne Agentis is cited, so people can see the sources behind an answer, and output is delivered in familiar formats such as Word, PDF, PowerPoint and Excel.

06Governance and standards

Security and governance are built in, not added on:

• Role-based access, so the right people see the right data.
• Metering, so usage is measured and visible.
• An audit trail, so actions are recorded and traceable.

These controls are mapped to ISO 27001, the GDPR and the EU AI Act.

07Partners and suppliers

Partners in the FederatedOne EcoPartner network build on a governed platform, under the same approval gates and audit trail. We expect the companies we work with to apply standards consistent with this policy.

08Reporting concerns

If you have a concern about how AI is used in a FederatedOne component or by us, contact us at info@federated.one or through our contact page. We will review every concern and respond.

09Changes to this policy

We will review this policy as technology, regulation and our services evolve. The date at the top of the page shows when it was last updated.