Risk oversight, regulatory governance and the assurance of models and AI

A risk committee is only as good as the numbers and scenarios put in front of it. She has built stress-testing and capital-reporting cycles from the inside, so she knows where the assumptions sit and which ones a committee should press on.

Scope: Stress testing (EBA, PRA, CCAR) · ICAAP and capital reporting · Market and credit risk · RWA and capital allocation

Her regulatory depth was earned in delivery rather than advisory, under the US Federal Reserve, the PRA, the EBA, the HKMA and the MAS. For an Indian board facing RBI scrutiny, that is experience of what a supervisor actually tests.

Scope: Basel III and CRR3 · FRTB · EBA and PRA reporting · Dodd-Frank, MiFID II, EMIR, REMIT

Boards approve models they cannot inspect. She holds the Certificate in Quantitative Finance and has run model calibration and validation to supervisory standard: she can read a validation report and ask what it does not say.

Scope: Model validation and calibration · VaR, SVaR, FRTB · Back-testing · SR 11-7 principles

A board that attests to regulatory numbers is relying on data lineage it rarely sees. Her BCBS 239 work is the audit committee's data-integrity mandate, done from the inside.

Scope: BCBS 239 lineage · Data-quality controls · Regulatory reporting · ESG and CSRD data

The question now reaching Indian boards is how to adopt AI inside a regulated function without a control gap internal audit would reject. She has answered it once in production.

Scope: AI governance in regulated functions · Controls for AI/ML in production · AI-enabled fraud risk · AI-enabled risk analytics