Financial services
Model risk management already exists here, and AI governance has to sit alongside it rather than duplicate it. Evidence from model registries and approval workflows carries straight into ISO 42001 and the EU AI Act.
Industries
What changes between sectors is which obligations bite first and where the evidence already lives. The control library, the crosswalk and the agents underneath do not change — which is why a second framework costs far less than the first.
Where it lands first
A rough guide to what tends to dominate the early programme by sector.
Model risk management already exists here, and AI governance has to sit alongside it rather than duplicate it. Evidence from model registries and approval workflows carries straight into ISO 42001 and the EU AI Act.
Pricing and underwriting models attract high-risk classification under the EU AI Act. Classification, impact assessment and human oversight are the obligations that arrive first.
Clinical and diagnostic AI carries both device regulation and AI-specific duties. Data quality, traceability and post-market monitoring dominate the control set.
Safety-adjacent AI in production systems needs a defensible record of what changed, when, and who approved it. Version history from your own registries becomes that record.
Usually certifying to ISO 27001 already. The security evidence you hold is reused to satisfy overlapping AI controls instead of being rebuilt in a second programme.
Transparency and accountability obligations arrive early and are scrutinised publicly. The traceable trail matters as much as the outcome.
What stays the same
Next / evaluate the platform
We will walk through the platform against your own governance and compliance needs, and what getting to audit-ready would take.