Tookitaki

An alert prioritization agent ranks work, the FinMate copilot summarizes alerts and drafts regulator ready reports, and network analytics surface relationships on top of model based detection. A rules engine with static thresholds and a case queue, the approach Tookitaki says it replaces, is all that would remain.

Each flagged transaction carries the data points and logic behind it, low code controls let compliance teams adjust thresholds, and rules can be simulated before deployment. FinMate drafts summaries and reports for an investigator to complete rather than filing them.

Drift detection and retraining are described as part of the model lifecycle, giving a validator a monitored process to examine. Reported false positive cuts carry no recall or false negative figure, and no validation package maps to a supervisory instrument.

United Overseas Bank recorded 5 percent more true positives and 70 percent fewer false positives on individual names in screening, with no period or baseline volume stated beside those figures. An unnamed Singapore digital bank deployed new scenarios 50 percent faster and cut threshold tuning and scenario testing effort by 70 percent, and Maya and DeeMoney are quoted. More than 100 billion transactions monitored and more than 200 million customer accounts are aggregates across the client base.

Hosting of shared model updates, their retention and lifecycle terms for contributed scenarios are not stated. Local training at each institution, with only model updates shared through a federated engine, answers the cross client question by name.

Sensitive customer data stays inside each institution even within the shared intelligence network, a product level privacy position. Retention schedules, subprocessors and processing terms for the data FinCense handles directly are unpublished.

Tookitaki's trust center lists SOC 2 and the Payment Card Industry Data Security Standard 4.0 as compliant, without the SOC 2 report type, auditor or period, and releases its security, data retention and access control policies on request. Encryption, multifactor authentication, strict access controls and monitoring around the clock are stated controls.

Pre configuration for the Monetary Authority of Singapore, Bangko Sentral ng Pilipinas and AUSTRAC, among others, builds scenarios, reporting formats and thresholds to those regimes. Alignment of that kind is not a register entry or supervised filing channel held by the vendor.

Independent validation under Singapore's AI Verify program, announced in September 2025, tested the systems on anti money laundering scenarios in English and Mandarin. False positive analysis across customer populations is unpublished, though mule and scam detection here falls heavily on migrant workers and remittance corridors.

Institutions can trace a held payment to the signal behind it and defend the decision to a supervisor. The customer whose payment was held is not told which system flagged it, and no accuracy commitment or indemnity accompanies the output.

The AFC Ecosystem is named as the origin of detection typologies, with more than 200 member institutions and over 400 typologies. Sanctions data providers, any model behind FinMate and a subprocessor list are not identified.

Monitoring, screening and case management are unified so a signal in one surfaces in the others, with real time coverage of Australia's New Payments Platform and PayTo. Which core banking platforms it connects to is left unstated.

Region choices for the cloud option and the residency terms that go with them are not set out, but where FinCense runs is the institution's choice: on premises inside its own infrastructure, in the cloud on AWS or Google Cloud, which most clients use, or as a hybrid of the two. Each of its markets applies its own outsourcing rules to where processing occurs, and the on premises route keeps that processing in house.

Terms for a bank taking only transaction monitoring rather than the full module set are not described, and FinCense carries no published price or basis of charge. Membership of the AFC Ecosystem is free and invite only.

Banks, digital banks and remittance firms are among more than 30 Asia Pacific clients, with case studies for a Singapore bank, a digital bank and an e wallet. Coverage runs deep in one region rather than broad across the world.