For AI labs
RL environments to train frontier models & evaluate compliance agents.
Realistic financial crime work: onboarding customers, monitoring transactions, screening sanctions and preventing fraud.
Action · Review transactions
Four environments
Built the way a financial crime team works its queue. An alert or onboarding file comes in; the agent gathers the evidence, applies policy and risk appetite, and decides whether to clear, escalate, request information or report. The rationale is judged as closely as the decision, the way a QA reviewer or examiner would.
-
KYC
Perform CDD and EDD: document requirements by customer type, product, jurisdiction and geography; collection and verification of the customer, related parties and beneficial owners; PEP, sanctions and adverse-media screening; customer risk rating.
-
Transaction Monitoring
Alert review against expected activity and source of funds and wealth; transaction and counterparty analysis across fiat and crypto; typology identification; closure or escalation to SAR/STR.
-
Sanctions Screening
Name, entity and wallet screening; true-match vs false-positive adjudication; ownership and control; payment routing; release, block, reject or escalate.
-
Fraud Prevention
Account takeover, APP scams, mule accounts and first-party fraud across cards and payments; device, behaviour and payment-history review; hold, release, customer contact or exit.
Benchmarks
FinCrime Bench grades agents on sealed, held-out cases in four tracks: KYC, transaction monitoring, sanctions and fraud. A decision counts only with the evidence behind it. Coming soon.
Working on agents for risk and compliance? We’re running private pilots with AI labs.
Bring your own model; we supply the synthetic cases, the policies and the grading.