AML training a regulator would respect: beyond the annual click-through

Every regulated firm runs anti-money-laundering training, and almost all of it is the same annual click-through: a generic module, a ten-question quiz, a certificate. Examiners have stopped being impressed — and for a good reason. That artefact proves attendance, not competence. It proves someone was logged in, not that they’d catch a real red flag on a Tuesday afternoon.

The annual click-through problem

The standard AML course is bought off the shelf, delivered identically to everyone from the teller to the BSA officer, and passed by scrubbing to the end and picking the recognizable answers. It satisfies the checkbox. It does very little for the actual objective, which is that a person confronted with structuring, an unusual wire, or a nervous walk-in knows what to do and does it.

The gap between “completed the AML module” and “can act on a red flag” is exactly the gap examiners have grown wary of — and it’s the gap a recognition quiz is structurally unable to close.

What examiners actually look for

US supervisory guidance is not subtle about this. The FFIEC BSA/AML examination manual expects training to be tailored to the specific responsibilities of each role and to reflect the institution’s own risk profile and procedures. Inadequate or generic training programs show up by name in enforcement actions. Examiners are looking for evidence of tailoring and evidence of understanding — not a spreadsheet of completion dates.

A completion log answers “who attended?” The question underneath the exam is “would this person actually recognize and escalate the thing our policy is designed to catch?” Those are different questions, and only one of them keeps you out of a finding.

Tailored means your policy, not a template with your logo

“Customized” AML content usually means a vendor template with your name dropped on the cover. Real tailoring means the training is grounded in your own BSA/AML policy — your risk appetite, your typologies, your escalation path, your SAR procedures — because that is what the employee actually has to apply. Attaching a PDF of your policy to a generic course is not the same as training on it.

This is where AI facilitation changes the shape of the problem. Instead of playing a fixed module, the AI runs a scenario and asks the employee to reason through it, drawing every prompt from the institution’s own procedures and citing the relevant section inline:

Mastery, not attendance

The second half of what an examiner wants is proof of understanding — and that is a measurement problem generic courses solve with a recognition quiz that anyone can pass. Facilitation replaces it with mastery: the learner has to apply the policy to a scenario and reason it out, and completion is gated on demonstrating that they can. What you get out the other side is an auditable record tied to the specific competency and policy section, not a checkbox.

What it looks like in practice

With ThinkPair, the workflow is short: upload your BSA/AML policy and procedures, and the facilitator runs scenario-based Socratic sessions grounded in them — asking, pressing, and citing the relevant section as it goes. Completion is scored by competency, and the record exports into the LMS you already run, so the audit trail lands where your examiners already look.

We’re pre-pilot, so we won’t claim an examination outcome we haven’t earned. But of every vertical we looked at, banking and financial-crime compliance is where the gap is widest: the regulatory expectation is already “tailored, and prove they understood,” and almost nobody delivers AML training that way. That is precisely the shape of training a facilitator is built to run — and precisely the shape a click-through never could.