Azentio Software

Core banking, treasury and insurance administration would all keep running without models, since they process transactions and keep records. AMLOCK detects on a scenario library and rule builder, with machine learning added beside the rules in its 2025 relaunch.

Alerts are risk scored, routed by branch, scenario and customer group to analysts, and worked through investigation and regulatory reporting, so a person stands before any filing. Suppression runs on rules and analytics, with no confidence threshold or share of alerts closed automatically published.

False positives cut by up to 40 percent is a ceiling with no baseline, method, institution or period, and with no true positive retention it cannot be read alone. Validation, recall and drift monitoring are unpublished for screening and identity models.

Boubyan Bank, Tamam Financing and Philippine National Bank are quoted with qualitative endorsements of their deployments. An anti money laundering deployment that cut manual effort is credited only to an unnamed Indian fintech.

Pattern detection across many institutions in overlapping markets benefits from pooled typologies, which would be valuable to the vendor and invisible to any single bank. Exclusion of client data from training, and separation between institutions, are not stated.

Digital onboarding collects identity documents and facial templates, which several of its markets treat as special category data, yet nothing on retention or purge on termination is published.

Certifications, SOC reports and penetration testing statements appear nowhere on the homepage or AMLOCK page. That gap is striking for a supplier running core banking and processing biometric data for banks on three continents.

A 2023 certification from the Accounting and Auditing Organization for Islamic Financial Institutions covers the Shariah compliant products and profit calculation engine, a real assessment that is not regulatory standing. As a software supplier the firm states no license or supervisory relationship of its own.

AMLOCK is said to assess risk on behavior, not bias, with an auditable rationale per alert, a design claim with no evaluation behind it. Face matching in onboarding across Asian, Middle Eastern and African markets has no published demographic performance.

Investigation and reporting workflows in AMLOCK give an institution a record to examine a screening outcome. False matches, wrongful liveness rejections and bad credit decisions have nothing standing behind them, and the affected person is not told automated processing was involved.

Machine learning in AMLOCK, face matching and liveness detection are described by function only, with no model, provider or build or buy position given. Biometric matching is often licensed, which makes the omission material.

iMAL is the system of record for the banks that run it, with lending, treasury and wealth modules attached natively and Islamic profit accruals and distributions computed inside the accounting core. The digital layer deploys onto the firm's own core or a third party core, though no third party core is named as certified.

Cloud ready architecture and AMLOCK support for Windows and Unix environments, AIX among them, imply installable software, but on premises, managed cloud or hybrid is never stated. No cloud provider, region or residency commitment is named for banks under national localization rules.

License basis, metering and implementation fees are unpublished across core banking, financial crime and ERP, and every route is an inquiry or demonstration request. A 10 to 12 week implementation for AMLOCK helps plan a project but not price it.

Each named segment carries a product line: ONEBanking for conventional and Islamic banks, policy administration for insurers, and wealth and capital markets products. Mid tier to large institutions across Asia, the Middle East and Africa are served from research centers in India and Lebanon.