SAS

AI powered detection, network analytics and entity resolution are genuine statistical work, yet scenario authoring, investigation workflow and regulatory reporting would all keep running without them. Institutions ran that surviving platform through two decades of anti money laundering practice.

Investigators sit at the center, with SAS Visual Investigator as the alerts interface and dynamic due diligence checklists as core components. Which alerts close automatically, what threshold routes a case to a person and how automated customer scoring is reviewed are not described.

Visibility into the decisions of AI and machine learning models is stated, and scenario logic is authored in a no code interface an institution can inspect. A validation package and mapping to SR 11 7 or OCC 2011 12 with backtesting evidence are not published.

Landsbankinn cut false positives by 90 percent, with daily false positive alerts falling from about 1,000 to about 100 within months, and its head of risk solutions, Skuli Jensson, is quoted. Analyst placements describe recognition rather than measurement.

Whether models tuned on one institution's alert outcomes serve another, whether customer data enters any shared model and what retention applies to training data are not stated. Security and compliance assurances on the managed cloud page leave that cross client question open.

Transaction level activity and customer risk profiles across many jurisdictions concentrate in one environment by design, yet no data protection agreement, retention schedule or deletion commitment for the portfolio is published. Counterparties with no relationship to the institution are profiled too, and nothing addresses them.

ISO 27001, SOC attestations from the American Institute of CPAs and FedRAMP authorization are named for the managed cloud rather than the financial crime line. Report type, certificate scope and dates are absent.

Monitoring for money laundering, enhanced due diligence and beneficial owner identification are sold as functions that exist because a rule requires them. Specific instruments, supervisors and filing channels are not named, and as a technology supplier SAS claims no license for this line.

Watchlist screening errors vary with naming convention, transliteration and script, and a false match means a frozen or closed account, while customer risk rating decides who is treated as higher risk. Broader trustworthy AI publications support scrutiny, but per population accuracy and disparate outcome testing for this product are unpublished.

Explainability for auditors and examiners lets an institution reconstruct why an alert fired or an entity matched, defend it and reverse it internally. Customers wrongly matched or scored as higher risk are not told which system produced the outcome, and no guarantee or accuracy commitment is published.

Detection models, the SAS Viya platform and managed cloud operations are all built in house, which closes much of the third party model question. Sanctions and watchlist data providers, and any third party model behind generative features, are not named.

Shared data, models and investigation workflows across fraud and money laundering teams form the integration argument, replacing separate systems with one platform. Core banking platforms, sanctions data providers and external case management systems are not named as targets.

Remote managed services let an institution keep its data on premises or in a private cloud, while the managed cloud runs on Azure, Azure Government or AWS. Regions hosting a given customer are not stated publicly.

Financial crime modules carry no published price, ladder or charging basis, and contact is required to learn any of it. SAS Viya and managed cloud services sit beside the industry solutions, which leaves total cost of ownership hard to see.

Banks, credit unions and other financial institutions are served, with Bangkok Bank and the French banking as a service provider Treezor among customer stories. Payments fraud, sanctions screening and regulatory reporting run as one maintained industry business under its own leadership.