AI that behaves like an experienced compliance professional.

No black boxes, no invented obligations. Every AI output in Mortevo Audit cites the specific rule or document it depends on, or it doesn't ship.

Our principles

Evidence-based. Every conclusion cites the rule, section or document it depends on. If it can't cite, it doesn't answer.
Uncertainty is stated. When the AI is not sure, it says so and points to the guidance a human should consult.
No invention. The AI is grounded in the Regulatory Knowledge Engine. It will not fabricate rule numbers, obligations or timelines.
Human sign-off. AI output is a draft. Findings, remediation plans and outcomes are always confirmed by a person.

Where AI helps

Assisted findings. Given a reviewer's note, the AI suggests the specific MCOB or Consumer Duty rule likely breached and drafts finding language a compliance manager can edit.
Rule explanations. Plain-English summaries of complex rules, with side-by-side citations to the Handbook section they come from.
Remediation drafts. Prefilled remediation templates based on the finding category, editable before sign-off.
Regulatory change impact. When the engine ingests a policy statement, the AI drafts a short "what this means for your firm" summary, always linked back to the source.

Where AI does not act

Confirming a review outcome. AAL2 (multi-factor) is required and only a human can sign.
Closing findings. Closure is a human decision with recorded justification.
Creating or releasing legal holds. Separate human permissions apply.
Modifying the audit log, snapshot integrity, or role assignments.

Model choice and data handling

Mortevo Audit uses the Lovable AI Gateway to route requests to appropriate models. AI providers are contracted sub-processors and are listed in our Privacy Policy. AI prompts and completions are logged in the audit trail so that any output which materially influenced a review can be reproduced later.

No customer content is used to train third-party models. Requests are scoped to a single review or rule and stripped of identifiers the AI does not need.

Explainability

Every AI output in the app has an "explain" affordance which shows the citations, the confidence signal, and the prompt template used. Compliance and audit users can inspect this at any point without leaving the workspace.

Reporting an issue with AI output

If the AI produces something incorrect, misleading or unsafe, email compliance@mortivoaudit.co.uk with the review ID and a short description. We treat AI defects as a compliance issue, not a bug report.