Our position

Four principles for enterprise AI

This page sets out our approach. Architecture, governance, sovereignty and agent rights are explained on dedicated pages.

Control

Technology should constrain decisions rather than obscure accountability.

Principle

Evidence

Public statements and product status are based on verifiable results.

Principle

Independence

Architecture decisions are not tied to a single model provider.

Principle

Accountability

Decisions requiring subject-matter or legal judgment remain with people.

Principle

Who we work with

For organisations with stringent requirements

The right starting point depends on the data landscape, regulation and accountability model.

Data-intensive

Complex information landscapes require clearly defined responsibilities.

Organisation

Regulated

Review, evidence and oversight need to be considered from the outset.

Governance

Accountable

People retain authority over approvals and subject-matter decisions.

Operations