AI use statement

In client systems

• Every capability is scoped to approved data sources agreed with the client.
• Every consequential output has a named human reviewer identified by role.
• Inputs, outputs, reviewers and overrides are logged with client-defined retention.
• Capabilities are evaluated before production and re-evaluated on a defined cadence.
• Any operator can halt an automated path without seeking permission.

What we will not deploy

• Autonomous agents that take consequential action without human approval.
• Capabilities that generate client-facing output with no review point.
• Systems that draw on data sources the client has not approved.
• Models trained on client data without explicit written agreement.

In our own work

We use AI tools for research synthesis, drafting and analysis support. A Bridgelytic person reviews and is accountable for every deliverable, recommendation and client communication. Confidential client material is not entered into tools outside the boundary agreed with that client.

Website content

Content on this site is written and reviewed by Bridgelytic. Where a page describes a capability rather than a completed client engagement, it says so on the page.

Model providers

Where a third-party model provider is used in a client system, the data boundary, retention terms and processing location are documented and approved before the capability goes live.

Limits of this statement

This is a description of Bridgelytic's practices maintained by Bridgelytic. It is not a certification, an independent audit, or a compliance attestation under any framework.

Contact

Questions about AI governance: security@bridgelytic.capital.