Fraud Analytic & AML Solution Technology
An enterprise AML platform built and supported by a Thai team with more than a decade in fraud & compliance — engineered to fully align with Thailand’s anti-money laundering regulation from day one.
Solution Overview & Value Proposition
FAAST (Fraud Analytic & AML Solution Technology) is an enterprise-grade platform for real-time fraud detection, prevention, monitoring, investigation and reporting. It continuously monitors transactions and user activity across all delivery channels, analyzes them for high-risk or suspicious behavior, automatically intervenes by halting risky transactions, and alerts the relevant fraud, security and IT teams to investigate and act in line with the Bank’s own procedures.
The platform combines FAAST Self-Configurable Business Rules with the NerveML Engine, orchestrated through a flexible Visual Workflow Canvas. New detection rules can be deployed very quickly, enabling the Bank’s fraud team to keep pace with rapidly changing fraudster schemes, while the platform supports high-throughput real-time integration and batch processing.
How FAAST Works
FAAST operates as a continuous pipeline from data ingestion, through rule-based and behavioral analysis, to alerting and case investigation giving the Bank full visibility and control at every stage of the fraud lifecycle.
Step 1 — Integration
FAAST supports real-time, near real-time and batch integration, so each data source can be onboarded using the method that suits it best. FAAST’s specialists work with the Bank to identify the optimal integration point and technique for every channel.
FAAST provides flexible integration capabilities across real-time, near real-time, and batch environments. Our specialists collaborate with the Bank to identify the most effective integration strategy for each channel, maximizing data availability while minimizing implementation complexity
Step 2 — Analytics
Once ingested, every transaction and activity is evaluated by two complementary techniques:
- Rule-Based Analysis — compares activity against pre-configured rules one of the most effective ways to catch fraud patterns that have already been identified.
- Behavior Analysis — profiles each customer’s normal activity and flags meaningful deviations, for example an unusually large transfer, a transaction from an unrecognized device, or a payment to a never-before-used beneficiary.
Step 3 — Investigation & Case Management
When a suspicious transaction is detected, FAAST raises an alert in Case Management, the web application used by fraud analysts, anti-fraud managers and IT administrators.
Case Management presents full supporting detail so the investigator can decide whether to cancel the transaction, temporarily suspend the user or account, or release the transaction if it is legitimate every action captured in the case record for audit, following configurable workflow and approval steps that mirror the Bank’s existing maker-checker procedure.
FAAST Data Analytics — Fully Customizable Multi-Layer Data Analysis
FAAST’s analytic engine is built as a multi-layer data pipeline. Each layer is independently configurable, so the Bank can combine deterministic rules, behavioral profiling, risk scoring and extensible analysis in whatever sequence and depth best fits its risk appetite, data volumes and regulatory obligations without re-architecting the platform.
- Rules — best suited to known fraud patterns and regulatory compliance. Fast, auditable first line of detection.
- Behavior Profiles — establishes a behavioral baseline for each customer or account and flags statistical deviations, catching anomalies that static rules alone would miss.
- Risk Scoring — aggregates signals from the Rules and Behavior Profile layers into a single, prioritized score improving analyst prioritization, efficiency and automation.
- Additional Analysis — an open, extensible layer for further analysis machine learning models via the Nerve Platform, Watchlist, or external interfaces and data sources extending coverage without changing the core pipeline.
- Alerts — the pipeline’s output: a risk-ranked, evidence-backed alert routed into Case Management for investigation and decisioning.
Core Platform: FAAST Platform & NerveML Platform
FAAST is built on two integrated platforms: the FAAST Platform, which delivers self-configurable rules, case management and reporting; and the NerveML Platform, the machine learning engine that extends detection beyond rules
FAAST Platform
The FAAST Platform gives fraud analysts, monitoring teams and compliance officers the tools to detect, investigate and report on suspicious activity without developer support:
- FAAST Rule — a self-configurable rule engine: analysts create and modify detection logic directly, with no developer, release cycle or vendor change request required. Rule outcomes drive automated, fully configurable workflow actions hold, reject or block a transaction, suspend a user or account, or route the item for human review so risk is acted on in real time, not merely reported.
Complete rule coverage without source code changes • The library of node types provided with FAAST is designed to express majority of the fraud detection rules a bank requires — single-transaction evaluation, period aggregation and velocity, statistical behavior deviation, list matching, value comparison, incident raising, external service calls and reporting. • Because these node types are combined through configuration, new and revised rules are built without any change to source code — no custom development, no recompilation, no software release. • Rule changes therefore fall entirely within the Bank’s own control, rather than depending on a vendor development and deployment cycle. |
- Watchlist (Active List Management) — special lists used to calculate baseline scores for incidents and alerts against specific entities such as accounts and employees, matched using the FuzzyEngine for phonetic and spelling-variant detection.
- Rule Scoring Management — assigns configurable scores to different alerts, helping analysts prioritize and focus on the risk that matters most.
- Alert Management — notifies analysts of detected suspicious activity, with all relevant information attached for investigation.
- Case Management — a workspace where confirmed-fraud alerts are investigated; related alerts can be grouped into a single case. Role-based investigator actions cancel a transaction, suspend a user or account, or release a legitimate transaction are captured in the case record for audit.
- Workflow Management — the Visual Workflow Canvas (FlowNode UI) lets the Bank design case routing, escalation and approval steps that mirror its existing maker-checker procedure configurable graphically, without code.
- Reports — the self-service Report Wizard lets analysts and managers build operational, management and regulatory reports independently, based on a business entity or a template, without developer support.
NerveML Platform
The NerveML Platform (NerveML) is FAAST’s machine learning core. It runs alongside rule-based detection, so coverage is not limited to fraud patterns that have already been built into rules.
NerveML analyzes behavior profiles of each customer’s normal activity and flags meaningful deviations, an unusually large transfer, a transaction from an unrecognized device, a payment to a never-before-used beneficiary, or an impossible-travel login sequence. Rule-based and ML detection run in parallel: rules catch known fraud patterns efficiently while ML models extend coverage to emerging threats and progressively reduce false positives.
FAAST Solution Package
Anti-Fraud Solutions
Detection and prevention of external and internal fraud across every transaction channel, protecting both customers and the organization itself.
- Transaction Fraud — detects patterns such as transfers to suspicious or newly created accounts, unusual activity on otherwise dormant accounts, rapid in-and-out (clear-out) movement of funds, many-to-few fan-in transfers, and activity affecting vulnerable customer segments.
- Internal Fraud — monitors employee and teller activity for indicators such as unauthorized withdrawals from dormant accounts, excessive account inquiries, unusual volumes of error corrections, or system access outside normal working hours.
- Application Fraud — screens loan and account applications for indicators such as watchlisted applicants, guarantors linked to multiple suspicious applications, income inconsistent with the application, or loan proceeds transferred immediately to high-risk accounts.
- Case Management & Investigation — every alert is investigated and actioned through the same Case Management workspace, with role-based investigator decisions captured for audit.
Anti-Money Laundering (AML) Solutions
Monitoring, screening and reporting aligned to regulatory requirements, giving compliance and AML teams end-to-end control from detection through to filing.
- Customer Due Diligence — the Bank configures its own risk factors country, customer type, product, delivery channel and transaction behavior and sets the risk level and weighting of each; FAAST then evaluates every customer against that model and assigns a risk rating. Ratings are maintained on an ongoing basis: periodic review cycles run automatically by risk band, and event-driven re-rating is triggered when a customer’s profile or transaction behavior changes so the rating reflects current risk rather than onboarding-day risk.
- Sanction & Watchlist Screening — real-time and batch screening of customers, counterparties and transactions against sanction, PEP and internal watchlists, using the FuzzyEngine for fuzzy and identical matching across names, entities, vessels, goods and countries. Screening runs at onboarding and on every transaction, and whenever a list is updated FAAST automatically rescreens the existing customer base against the changed entries so a newly designated party is detected immediately, without waiting for that customer’s next transaction or review date.
- Suspicious Transaction Monitoring — continuous monitoring against configurable AML typologies, including structuring/smurfing, mule account activity, rapid in-out (clear-out) movement, activity inconsistent with customer profile, and high-risk counterparties or jurisdictions.
- Alert Investigation & Adjudication — analysts review and close each alert as a false positive or confirmed suspicious activity within Case Management, keeping compliance teams in full control of every decision.
- Regulatory Reporting — automated generation of suspicious transaction reports pre-populated from case data, with configurable templates matching each regulator’s required format; teams verify, edit and submit within required filing timelines, or create a manual report for a transaction the system did not flag.
Key Benefits — FAAST vs. Legacy Fraud Systems
FAAST is available and supported on IBM LinuxOne
