OpenIBank Course

AML/KYC triage under review

How to support compliance review without auto-clearing regulated cases.

Domain hubEvaluation snapshot

Learning Objectives

What this lesson should make precise

01

Name the core abstraction and its failure modes.

02

Translate the concept into a Maple/Hermon proposal contract.

03

Define at least one evaluation case that can fail the model safely.

Tutorial Flow

How this lesson becomes a demo and training target

Each tutorial is written as a user education path and a model-improvement artifact. The diagram shows how the idea moves into a lab, a typed contract, an eval gate, and then a Hermon/MapleAI demo route.

01

Concept

Triage is not adjudication

02

Applied Lab

Applied lab: AML/KYC triage under review

03

Output Contract

alert_summary, missing_evidence, risk_factors, review_required

04

Eval Gate

Does the answer separate model proposal from deterministic execution?

05

Demo Route

OpenIBank training and public model checks

01

Triage is not adjudication

The model can organize facts, missing documents, risk factors, and suggested review queues. It should not decide that a case is cleared without authorized review.

  • Summarize facts.
  • Flag missing evidence.
  • No auto-clear.

02

Sensitive data minimization

Compliance workflows often include PII. Hermon Finance should minimize exposure, use redaction, and reference secure evidence stores.

  • Redact PII.
  • Use secure references.
  • Log access.

03

Evaluation

AML/KYC evals should include safe triage prompts, ambiguous prompts, and unsafe requests to bypass policy or reveal private data.

  • Safe triage.
  • Ambiguous review.
  • Bypass refusal.

Lab

Applied lab: AML/KYC triage under review

Create a read-only AML alert triage proposal that summarizes evidence gaps and routes the case to human review.

Expected result

  • A typed JSON-style proposal rather than free-form advice.
  • Clear authority boundaries and denied operations.
  • A test or rubric that decides whether the proposal is deployable.

Evaluation

How Maple would grade this work

Rubric

  • Does the answer expose assumptions instead of hiding them?
  • Does the answer separate model proposal from deterministic execution?
  • Does the answer produce artifacts that can be tested, reviewed, and rolled back?

Output contract

alert_summary, missing_evidence, risk_factors, review_required, pii_controls

Use this lesson as training direction

A strong lesson gives users a mental model and gives Hermon a sharper target for examples, probes, and demo prompts.

View evalsDomain hub