Every session ends with something the audience can act on Monday — questions to ask, evidence to demand, and a way to test their own posture.
Building AI governance frameworks that survive the regulator — accountability across the three lines of defence, and the evidence an examination requires.
The institution reported 14 AI systems. Discovery found 41. Why inventory — not intent — is the first examiner question, and what a vendor-embedded sweep actually surfaces.
45–60 min · board or executive committeeThe architecture of examination-durable AI governance in Fed, OCC and NCUA regulated environments. Five stages, one gate, and why the weakest link sets the posture.
45–60 min · risk and auditDecision authority in agentic systems. The accountability gap that opens when an agent acts against money without a mapped human owner, and how to close it before deployment.
30–45 min · technology and riskExamination cycles, vendor change notifications and model drift set the rhythm. What an annual assurance cadence produces, and why it does not require a new decision each year.
30–45 min · boardAsk for the AI inventory. Then ask who built it, and whether vendor-embedded features are on it.
Every system needs one accountable person. Not a team, not a committee. A name.
Autonomy times blast radius. An agent that executes against money is not a chatbot.
Posture is set by the lowest score, not the average. Find the lowest one.
Speaker biography, headshot, talk descriptions and technical requirements.
Speaker bio & media kit → Enquire about a date →