TrustDecision

The removal test leaves rule engines, which is explicitly what the company positions against. Machine learning and deep learning drive fraud and credit scoring, graph based models identify fraud rings and collusion by connecting users, devices and transaction behaviour, behavioural and device signals detect automation before entry, and alternative credit scores are constructed for applicants with no conventional file. Federated learning trains across institutions without centralising their data, and decisions return within twenty milliseconds.

Control sits with the institution's own risk team by design. No-code tools let them deploy rules, simulate decisions and compare outcomes in real time before anything goes live, which is experimental control rather than vendor configuration, and domain experts work alongside clients during proof of concept to refine strategies. Every action carries an explainable reason for audit.

Against that, the in-repayment framework recommends and can trigger consequential actions including limit freezes and collections prioritisation, and nothing describes what executes automatically versus what a person authorises.

Two controls stand out. Every decision carries an explainable reason, which the company ties directly to compliance and audit readiness rather than offering it as a feature, and risk teams can simulate decisions and compare outcomes before deploying a change, which is pre-deployment testing owned by the customer rather than validation asserted by the vendor. Latency is published at twenty milliseconds.

Held at B because the only effectiveness figures, detection lift against legacy rule engines, come from secondary analysis rather than the company, and no accuracy, false positive rate or validation result is published directly.

A major Philippine bank launched a fraud risk management and anti money laundering platform on this technology in December 2025, with its chief innovations officer quoted by name, which is recent and specific. The company operates from Singapore with offices in Malaysia, Indonesia and the Philippines, and carries the backing of an established parent group in its home market.

Much larger figures circulate for the parent, covering client counts, signal volumes and market share, and they appear only in low authority secondary analysis rather than company or press material, so they are noted rather than relied on.

This is a direct architectural answer to the question left open at almost every vendor in this index. Federated learning means the collective intelligence that makes fraud detection work is built without pooling customer data centrally, so an institution contributes model improvement rather than records, and the company presents this specifically as the mechanism enabling shared defence across clients who cannot share data.

Held at B because the description comes from secondary analysis rather than the company's own published material, and no boundary policy, opt out or statement of what a client contributes and receives is set out formally.

The architecture addresses privacy structurally rather than by policy, running federated learning so that models improve across institutions without their data leaving their own environment, which the company frames explicitly as allowing shared intelligence without breaching data residency requirements. Privacy compliance is named as an accumulated capability alongside cloud native and multi cloud architecture.

Held at B because no data processing agreement, retention schedule or subprocessor register was located, and the platform ingests device, behavioural and transaction data on individuals across many jurisdictions.

No attestation, certification, trust centre or enumerated framework was located. Terminal security and threat intelligence analysis are named as technical capabilities, which describes what the product does rather than how the company is assessed. Large regulated banks running the platform on their own infrastructure will have conducted demanding security review, and none of that documentation is published.

A named statute anchors this, with the Philippine bank deployment described as aligning with that country's 2024 anti financial account scamming legislation, which is the specific law driving fraud investment in that market. Know your customer and anti money laundering obligations are named as covered functions, audit readiness is built into decision output, and the company positions compliance as one of its three core areas alongside fraud and credit. Held at B because no regulator is named for the platform itself across the ten or more countries served, each with distinct rules on automated credit decisions and data use.

The inclusion mechanism is specified rather than asserted, which is rarer than it should be. Alternative credit scores are built for thin file and new to credit applicants, and the company names how risk is managed while lending to them, through segment based strategies such as step up lending and low initial limits, so exposure grows with demonstrated behaviour rather than being refused outright.

A bank customer independently frames the outcome as faster and more inclusive access to credit. Held at B because device and behavioural signals carry the proxy risk recorded elsewhere in this index, automated limit freezes act against people already under strain, and no fairness testing or outcome analysis by segment is published.

No guarantee, indemnity or correction process was located. Explainable reasons serve the institution's auditors rather than the affected person, and the exposure is concrete: an applicant declined on an alternative score, a customer whose credit limit is frozen by an in-repayment model, or someone flagged through graph analysis as connected to a fraud ring by shared device or network association. Nothing describes whether any of them learns why, or how a wrong association is contested and removed.

The intelligence source is described structurally as a federated network across the client base plus third party sources, and no individual data provider, bureau, telecommunications partner or identity vendor is named. The parent group relationship is disclosed and is itself a dependency worth understanding, since the technology was incubated there and the international brand carries its heritage. No model provider, hosting arrangement or subprocessor list appears.

Data is ingested through development kits and interfaces at both device and server level, and the platform unifies signals from users, devices, transactions and third party sources into one decisioning layer, with the stated purpose of modernising legacy systems rather than sitting beside them. A cross channel framework spans the digital banking estate. No named core banking, origination or case management system appears, and no developer documentation was located.

An on premise option is offered explicitly and described in the right terms, built within the institution's own infrastructure for maximum control, security and compliance with vendor support available, and aimed at large regulated institutions running their own fraud teams. Cloud native and multi cloud architecture covers the alternative.

Residency is treated as a design constraint rather than an afterthought, since the federated approach exists partly so intelligence can be shared without data crossing borders. Held at B because no region, provider or specific residency commitment is published.

No pricing, packaging or basis of charge was located. The platform spans fraud, credit and compliance across the customer lifecycle with both cloud and on premise delivery, which would ordinarily carry very different commercial terms, and nothing indicates whether charge follows decisions made, customers screened or institutional licence.

Buyer coverage spans retail banks, digital banks, consumer lenders, payment platforms and fintechs, and functional coverage is genuinely end to end: onboarding, identity verification, transaction monitoring, promotion abuse, credit assessment, in-repayment monitoring and collections prioritisation, with anti money laundering and know your customer handled in the same system. Geographic footprint covers Southeast Asia from four offices with presence stated across more than ten countries. Non financial sectors are also served, and the financial capability is specific rather than generic, which is what distinguishes this from a horizontal platform.