CoreCompliance

How CCAI works.

CCAI screens entities through a structured protocol. Each stage produces traceable output. The full path from intake to attestation is deterministic, auditable, and reproducible.

01

Intake

Entities enter via batch upload or API submission. Each submission receives a unique trace identifier and is bound to a tenant context.

The system detects file schema, classifies entity types (company, individual, vessel, address), normalizes input, and masks detected PII before processing.

Intake produces a normalized entity set. No resolution decisions are made at this stage.

02

Screening

Each entity passes through the resolution waterfall — a sequence of layers ordered by cost and complexity. An entity stops at the first layer that produces a confident resolution.

Each layer records its result, confidence score, execution time, and the rules and data it consulted. This per-layer trace becomes part of the evidence record.

LayerMethodCost
L0Garbage Detection$0
L1Deterministic Matching$0
L2Vector Similarity$0
L3Semantic Reasoning~$0.005
L4Human ReviewVariable
03

Evidence Assembly

Every screening decision produces a DecisionEventPackage containing: resolution path, rule binding (version and hash), data binding (watchlist snapshot), confidence basis, full lineage, and the configuration snapshot active at execution time.

The DecisionEventPackage is assembled after resolution completes. Once assembled, it is immutable — no field can be modified, appended, or removed.

04

Monitoring

Entities enrolled in continuous monitoring are automatically re-screened when watchlist data is updated. Delta alerts surface when a previously cleared entity now matches, or when confidence scores shift beyond configured thresholds.

Each re-screening follows the same protocol and produces the same evidence structure as the initial screen.

05

Attestation via Intelligent Analyst

After evidence assembly, the DecisionEventPackage is handed to Intelligent Analyst — the separate truth and attestation layer. IA records the decision in an append-only truth ledger and binds a VeracityReceipt (receipt_id, receipt_hash) to the decision.

CCAI produces the screening decision. Intelligent Analyst attests it. These are separate systems with separate data stores. Neither can modify the other’s records.

This separation ensures the evidence record is independently verifiable.