AI use & governance
Use AI where it improves the work. Keep responsibility human.
Novarra AI uses an evidence-led operating principle: AI can accelerate analysis and synthesis, but it does not remove confidentiality boundaries, source verification or human responsibility for material judgment.
Public principles
The same governance discipline applies to our own use of AI.
These are operating principles, not a claim that Novarra AI holds ISO 42001 or another AI-management certification.
Human accountability
AI may assist analysis, drafting or synthesis, but a human reviewer remains responsible for material conclusions and client-facing outputs.
Confidentiality boundary
Sensitive client, transaction or portfolio information should not be placed into unapproved public AI services.
Evidence before assertion
AI-generated analysis is treated as working material until it is checked against primary evidence or an appropriately identified source.
Tool approval
Permitted AI and analysis tools should be agreed for sensitive engagements together with access, retention and data-handling expectations.
Testing where material
Material AI systems should be evaluated for performance, failure modes, security, privacy and governance in the context in which they will be used.
Decision authority stays human
Novarra does not position AI output as an autonomous investment, board or management decision.
Client engagements
Sensitive AI use is an engagement-design decision.
Before sensitive evidence is processed, the engagement should define which AI tools are permitted, what data may be used, what must remain outside AI workflows, how outputs are reviewed and what evidence is retained for material conclusions.
Last reviewed: 3 October 2026.