Two commercially distinct assets. (1) Purpose-built labelled test corpora: synthetic transaction populations seeded with known money-laundering patterns (Red Flag Tests and transaction monitoring validator scenarios) and sanctions test records selected from named sanction lists with deliberately applied fuzzy-matching mutations, where the correct answer is known by construction — a labelled ground-truth evaluation set rather than an observed dataset. (2) A cross-institution system-performance panel: every customer's screening and monitoring results pass through one portal, so the operator holds comparable detection-rate, miss and false-positive profiles by screening or monitoring product, configuration, fuzzy-logic setting and filter regime — a vendor-comparison view no single bank or vendor can assemble. Access is already subscription-gated.
Equities in UK, US, Europe. Related tickers REL.L, WKA.AS, FICO, NSTB.L, INF.L, JPM, HSBA.L.
Tabular
Test corpora are described as large synthetic simulations of laundering patterns plus sanctions records drawn from multiple named lists, distributed as downloadable benchmark files — plausibly millions of synthetic records per corpus, but no record count or file size is published to confirm it
Labels are designed in rather than inferred, which makes them unusually clean: per-record suspicious versus non-suspicious ground truth, alert-versus-miss outcome per system under test, expected-versus-actual detection for each seeded laundering pattern, fuzzy-matching difficulty grades on mutated sanctions records designed to test name-matching capability, false-positive labels tied to specific filter configurations, and threshold-sensitivity labels showing at what parameter each pattern becomes detectable.
AML Analytics (AMLA / Analyser Online) — Two commercially distinct assets. (1) Purpose-built labelled test corpora: synthetic transaction populations seeded with known money-laundering patterns (Red Flag Tests and transaction monitoring validator scenarios) and sanctions test records selected from named sanction lists with deliberately applied fuzzy-matching mutations, where the correct answer is known by construction — a labelled ground-truth evaluation set rather than an observed dataset. (2) A cross-institution system-performance panel: every customer's screening and monitoring results pas…
Not yet on file: Sample · Data dictionary · Data frequency · Reporting lag · Point-in-time · License