AI Supervision for Compliance Teams
The second line’s core session. How to supervise AI systems you did not build and cannot see inside — built around the controls and risk your team owns.
Compliance, risk, model-risk and audit teams are now expected to govern AI systems the business has already bought or built — often without a clear framework for doing so. “AI governance” can sound abstract until a regulator asks how a specific model was approved, monitored and documented. This workshop turns it into concrete controls your team can actually run, using your real workflows rather than stock demos.
Heads of Compliance, MLROs, and compliance monitoring and assurance teams inside regulated financial firms.
Supervise the AI your business has already deployed — with a control framework, a risk register, and the questions that surface real exposure.
Supervisors have set out clear expectations for how financial institutions manage AI across its whole life cycle — in financial-regulator guidance, in long-standing model-risk supervision, and in the EU AI Act. The pressure now is to show a working framework, not a policy document that no one maintains.
- Describe what 'AI governance' means as concrete controls and records, not an abstract policy
- Find AI use across your business lines and grade each system by the risk it carries
- Set clear demands for third-party and embedded AI and verify vendor claims instead of trusting them
- Explain explainability, bias and model drift in terms your whole team can act on
- Document AI decisions so the record survives a regulator's questions later
- Stand up an AI risk register your team will actually keep up to date
A starting agenda — every session is shaped around your team, your tools and the risks you’re managing.
- What "AI governance" really means inside a regulated firm What “AI governance” means in practice inside a regulated firm — the people, controls and records a supervisor would expect to see, not a glossary.
- Spotting and grading AI risk across your business lines A simple method to find AI use across your business lines and grade each one by the risk it actually carries.
- Third-party and vendor AI: what to demand and how to check it What to demand from vendors and embedded third-party AI, and how to verify their claims instead of taking them on trust.
- Explainability, bias and model drift in plain terms Explainability, bias and model drift (when a model quietly gets worse over time) explained in plain terms your whole team can act on.
- Documenting AI decisions so they survive a regulator’s questions How to document AI decisions so the record survives a regulator’s questions months or years after the decision was made.
- Building a practical AI risk register your team will actually maintain Building an AI risk register your team will keep up to date, rather than a spreadsheet that’s abandoned after the audit.
Every team’s needs are different. We’re happy to talk it through and tailor the session to yours — let’s talk →
Bring "AI Supervision for Compliance Teams" to your team.
A short conversation about your team, your risk, and the session that would move them. No pitch deck — just the right scope and dates.
Enquire