On 17 September 2026 Intrepid Fox joined the AI innovation panel at the 6th Financial Innovation Forum — Payments & RegTech by QUBE Events. Founder Roman Zilber sat with bankers, lawyers and compliance professionals. The discussion turned out to be less about whether to use AI and more about how to supervise it.
Three things came out of the discussion.
AI in compliance is mainstream. Nobody treated it as experimental or new. The question has moved from "should we" to "how do we run this properly".
The use case is still augmentation. AI does the research, the analysis and the repetitive work. People stay responsible for judgement and for the decision. Intepid Fox KYB agents assemble the case. The compliance officer accepts or rejects it.
And then the hard one: how do you keep a human in the loop without making the human the new bottleneck?
"Human in the loop" is easy to write into a policy. It is much harder to do without rebuilding the queue you just automated. If agents produce ten times more completed cases and every one lands on the same desk for a full manual re-check, the constraint has not been removed.
The answer is not less oversight. It is oversight designed on purpose.
In practice that means three things. The reviewer should see evidence, not raw work: a structured file with every source attached, not a folder of documents to read again from scratch. The reviewer should focus where judgement is genuinely required, not at every step for the sake of the audit trail.
A review function that never overrides anything is not oversight. It is a rubber stamp with a person's name on it.
That last question stayed open in the room, and we think it is open across the industry. Nobody has settled how to measure whether human oversight of AI is actually working. It is worth settling before someone settles it for us.
