Concepts
The domain model
Nine ideas that the guides and the API reference both assume. Read them in order the first time; after that they are reference.
These pages describe how the platform models an AML programme — not how to operate it. For that, see the guides.
Tenancy and isolation
One deployment holds many institutions. Nothing about that arrangement is allowed to be the reason two institutions' data meet.
Parties and customers
A party is anyone the platform knows about. A customer is a relationship with one — not a different kind of thing.
Risk and due diligence
A risk rating is a number with a reason attached. The reason is what survives to the review, so the model records both.
Transaction monitoring
Activity arrives from two paths and is stored once. A rule is a window and a predicate, and every alert keeps the trace back to what produced it.
Screening and watchlists
Match on a party, score on evidence, decide on corroboration. A hit records what it matched and against which version of which list.
Alerts and cases
An alert is a signal. A case is the investigation of a subject. The workspace exists so an analyst never re-keys context that the system already holds.
Regulatory reporting
Reporting is a mapping layer with a swappable schema — and it refuses to lie about which schema it used.
Audit and evidence
Append-only is a claim about the application. Tamper-evident is a claim about anyone — including someone holding the database credentials.
Configuration and approvals
Configuration is a subsystem, not a settings table. Changes are proposed, reviewed and versioned, and the version in force at a decision is recoverable.