ThetaRay: financial crime detection, screening and AI investigation

Transaction Monitoring analyses relationships between activity, customers and external information to identify patterns across accounts and jurisdictions. The fact sheet describes domestic and cross-border monitoring, decision histories and auditable alerts. Named use cases include layering, structuring and money-mule activity. These are intended detection capabilities; the supplier’s descriptions do not establish a guaranteed detection rate for a particular customer’s data.[7]

Screening combines machine learning with deterministic matching and secondary attributes such as birth date, address and entity type. Customers can configure screening criteria and risk context. Its fact sheet also describes risk scoring using transaction history, geography and connections to higher-risk entities, with real-time alerts. Screening outputs therefore need a defined review and escalation process.[8][5]

Customer Risk Assessment uses KYC records, transaction information, external databases and adverse media to create a changing customer profile. It can categorise customers into risk tiers and update scores as activity changes. These scores support onboarding and ongoing review; the firm must decide which actions its policies require for each level of risk.[9]

RAY’s product page places its assistant in ThetaRay’s Investigation Center. It gathers case data, checks customer and counterparty activity against KYC information, reviews geographies and historical alerts, and organises a report. Analysts can use conversational questions, document summaries and controlled open-source research to examine the evidence and determine subsequent action.[4]

A current ThetaRay white paper describes RAY as an API-based layer for existing case-management environments. Its FX-institution example connects ThetaRay detection with Casmex and NICE Actimize. Compatibility, access permissions and connector work need confirmation for each customer.[11]

ThetaRay directs prospective customers to a discovery call. The reviewed product and sales pages do not publish a standard monthly subscription, per-transaction amount or setup charge. The commercial proposal needs to specify purchased solutions, usage volumes, implementation and support.[14]

ThetaRay describes its standard offering as SaaS. A deployment still needs a defined data scope, links between customer and payment records, agreed investigation responsibilities and acceptance testing. The quality and completeness of those inputs affect what the monitoring and risk models can observe. The public company page does not set one implementation duration for every customer.[2]

The Customer Risk Assessment page describes a sequence beginning with data exploration and the institution’s requirements. It then classifies customers, develops a semi-supervised model and uses weighted features to explain the assessment. Ongoing review happens through the Investigation Center. This makes risk categorisation and the institution’s review process part of the implementation, rather than simply switching on a score.[6]

The white paper also reports RAY running within a European banking group’s own private infrastructure, with data retained inside its security perimeter. That case progressed through real-alert scoping, tuning, user acceptance testing and phased production. It shows an enterprise delivery arrangement; it does not establish an off-the-shelf on-premises entitlement for every ThetaRay module.[11]

The public privacy policy describes safeguards, purpose-dependent retention and possible transfers outside the EEA under adequacy decisions or contractual protections. It identifies website and service information broadly, but does not specify one hosting region or a retention schedule for every institutional deployment. Those details should be established for the actual datasets and contracted service.[15]

Contracted support is governed by the SaaS service-level agreement; professional services require agreed scope. Public response times are not specified.[13]

The partner programme describes training and enablement for commercial and technical teams. That partner offer is separate from a customer’s support entitlement. Buyers should establish who owns operational incidents, integration changes and model questions, and how escalation works across any participating implementation partner.[10]

ThetaRay reports ISO 27001 certification and describes its monitoring results as explainable with audit trails. The public marketing page does not supply the certificate’s full scope or an assurance-report date. The Trust Center is the supplier’s route for further material; certification should be matched to the service and operating environment under consideration.[12]

An October 2025 announcement reports that Kaufman Rossin completed validation of the AML Transaction Monitoring and OFAC Screening modules in August 2025. It concerns those models’ accuracy, transparency and governance. The announcement describes plans to extend validation to other modules, so it should not be read as evidence that every later RAY feature or customer configuration received the same review.[17]

Screening and monitoring support an institution’s compliance processes. Model validation provides evidence about a specified system; it is not a banking, brokerage or payment-service authorisation, and does not determine whether a customer’s entire AML programme is effective. The purchased product, available data and operating controls must be assessed together.[17]

The current company page names Brad Levy as chief executive. ThetaRay’s privacy policy identifies ThetaRay Ltd. and an international headquarters in Hod HaSharon, Israel; its modern-slavery statement also identifies ThetaRay Inc. and affiliates. Current office contacts include New York, London, Madrid and Dubai.[2][15][18]

The customer’s order form identifies the contracting ThetaRay entity.[13]

ThetaRay acquired Screena in July 2024, bringing its cloud screening technology into the wider product suite. The acquisition announcement names JVP and Portage among the company’s principal investors. Screena is therefore part of this provider family, while those investment relationships do not make unrelated portfolio companies part of the offering.[16]

Earlier company material uses SONAR for the SaaS transaction-monitoring product and records the initial Screena partnership in April 2022. The current site presents Transaction Monitoring, Screening, Customer Risk Assessment and RAY as its principal product names. Historical SONAR references should be read in that product-line context.[19][1]

Operator

ThetaRay Ltd. and affiliated companies; contracting entity specified in the order form

Service scope

AML transaction monitoring, screening, customer risk assessment and investigation assistance.

Customer types

Banks; Fintechs; Payment-service providers; Financial institutions

Delivery

SaaS; API

Pricing publication

Quote only

Headquarters

Hod HaSharon, Israel

Platforms

Transaction Monitoring; Transaction and Customer Screening; Customer Risk Assessment; RAY

Integrations

REST APIs; Partner white-label API; Case-management integration scoped to the customer

Pricing scope

Discovery-led pricing; scope and charges are set in the commercial order.

Implementation

SaaS and API integration; a separate RAY customer case documents private-infrastructure deployment.

Support

Contracted SaaS support and separately agreed professional services.

Permissions and regulatory scope

Financial crime detection and investigation technology supporting institutional compliance.

Provider metadata checked 21 Oct 2026Source 2Source 16Source 14Source 1Source 10Source 15Source 5Source 12Source 19Source 13Source 11

ThetaRay’s public examples and marketing material report reductions in alerts and investigation effort. Those outcomes depend on the customer, dataset, configuration and comparison period. They are not universal service levels. A useful evaluation therefore measures both the workload produced and the quality of detected activity on the institution’s own representative data.[3]

The supplier’s current architecture paper describes feeding structured analyst outcomes back into detection and testing model changes before production. For an institution adopting that process, model-change approval and evidence retention need defined owners. An analyst’s case disposition can become an input to a changing control, making consistent recording and review consequential beyond the individual case.[11]