Stop fraud before it turns into a loss.
One ecosystem that detects in real time, understands with artificial intelligence and acts by reaching the customer in seconds. A transaction engine, cognitive AI and automatic validation via WhatsApp and voice, working as one team.
Detect and manage fraud from a single module.
- A rules engine your team builds and tunes
- Real-time transaction monitoring
- Alerts and cases in a single console
When you play it, YouTube may use its own cookies. Learn more.
Detect. Understand. Act.
FRAML MS AntiFraud brings together on one platform what usually takes three vendors: the decision engine, analytical intelligence and automatic customer contact.
FRAML MS AntiFraud modules
- Dynamic rules and AI/ML models trained on your own data
- Spending profile and habits for every customer
- Alert management with smart escalation
- No-code integration with Dynamic Message Mapping
Cerebrito
- Natural-language queries
- AI-suggested rule optimization
- Network risk graphs
- Device fingerprinting and mobile telemetry
Alert Defense
- Chat Defense via official WhatsApp
- Voice Defense with an AI voice agent
- Preventive block or immediate release
- Audio, transcript and full traceability
One platform, one integration, one view of risk: fraud and AML/CFT in the same system.
Everything your fraud prevention team needs, in a single console.
A real-time and near-real-time decision engine, multichannel (web, app, ATM, POS, call center) and multi-tenant, built so analysts can decide fast and with evidence.
Real-time management and operational visibility
Indicators that update with every transaction coming in through the API, and a customer record that shows how each customer normally behaves.
- Key metrics: today's alerts, critical risks, evaluated transactions and active monitors.
- Team management: cases closed per analyst and fraud prevented per rule.
- Spending habits: average amounts, channels, times of day and deviations from their segment.
- BI reports: custom SQL queries and CSV/XLSX export for corporate and regulatory reporting.
Management dashboardUpdated
Fraud prevented per ruleLast 12 days
Rules that adapt to your business, not the other way around
Risk analysts design, activate and govern rules without depending on IT, and manage every alert with full traceability.
- Dynamic rules engine: aggregates, velocity, channel combinations, geography, BINs and response codes.
- Alert inbox: block, confirm or dismiss in one click, with history and an analyst log.
- Escalation workflows: the most severe cases move up to supervisors or committees automatically.
- Exception lists: trusted IPs, blocklisted merchants and temporary exclusions (e.g., travel) to reduce friction.
Rules in production6 active
COUNT(tx) > 5 WHERE t < 10 mindevice_new AND amount > p95first_beneficiary AND channel = APPmerchant_id IN blacklistTest every rule on real data before you switch it on
Measure effectiveness and anticipate the impact on false positives against your history. Decide with numbers, not gut feeling.
- Rule simulator: test new conditions against historical transactions.
- Transaction and alert simulator: replay attack scenarios and validate the engine's response.
- Shadow mode: AI models run in parallel before taking over scoring.
Simulation · “APP Velocity” ruleBacktesting
Illustration for demonstration purposes. Results are calculated with each institution's own data.
Connect any channel without writing a line of code
Our integration edge: map and transform transaction messages in real time, and expose or consume services in minutes, with no deployments.
- Formats: ISO 8583, XML and JSON, with recursive reading of nested fields.
- REST and SOAP web services: authenticated GET/POST endpoints, ready in minutes.
- Databases, files, sockets and ETL: PostgreSQL, SQL Server, Oracle; fixed or variable length; batch or real time.
Connected sourcesNo code
Models trained on your institution's data
AI spots patterns that static thresholds miss, and graphs reveal the connections between accounts, merchants and branches.
- Supervised: Random Forest and Gradient Boosting (XGBoost, LightGBM) to prioritize alerts.
- Unsupervised: transactional outlier detection and distance-based patterns.
- Controlled lifecycle: precision, recall and false-positive metrics; continuous retraining.
- Correlation graphs: circular flows, bridge accounts and chained transfers.
Deployed modelsLifecycle
v4.2 · Recall 0.91 · AUC 0.95v2.0 · 14 clustersv1.3 · shadow testingcorrelation graphsIllustrative metrics.
Governance, traceability and control, audit-ready
Every action is logged and every user sees only what they should. The multi-tenant architecture fully isolates each institution's operation.
- Immutable audit trail: timestamp, IP, user, module and details of every change.
- Granular profiles: permissions by module, by button and for visibility of sensitive data.
- Access policies: password complexity and expiration, lockouts and sessions.
- Multi-tenant and business groups: rules, alerts and message mappings segregated by company.
Audit logImmutable
fraud.supervisor · 10.4.12.8analyst.03 · 10.4.12.21risk.admin · 10.4.12.2analyst.07 · no permission
Ask in plain language. Get answers from an expert analyst.
Cerebrito is the cognitive layer of FRAML MS. It analyzes your entire transaction universe in real time, talks with your team in natural language and tells you which rule to improve, which model to review and which network to isolate.
AI chat
“How many alerts were generated this week?” Instant summaries and explanations.
Forensic rule monitor
Efficiency of every rule plus AI suggestions: calibrate, relax or archive silent rules.
Model audit
Review of R and Python algorithms, data quality and projections before production.
Intelligence dashboard
Transaction universe, alerts, overall precision and active rules in a single view.
Organized fraud leaves traces. Cerebrito connects them.
A real-time graph engine maps accounts, devices and IPs. Instead of evaluating transactions in isolation, it reveals the whole network.
- Cluster detection: a single phone operating multiple identities, or fund-dispersion schemes (smurfing).
- Intelligence record per node with the risk level of every account or terminal.
- Immediate isolation of the suspicious network, in one click.
Shield your app and portal against impersonation and automated attacks
Canvas fingerprint, unique Device ID, IMEI, brand, model and screen resolution to detect spoofed devices.
Public and local IP, ISP, carrier, VPN and proxy usage to identify anonymization networks.
Emulator detection, GPS location and spoofed GPS (Mock GPS).
Proprietary technology + expert risk judgment.
With roots in fraud prevention and AML/CFT compliance going back to 2002, we support banks, fintechs, credit unions, trust companies and insurers in Colombia and Latin America.
Fraud and AML on one platform
FRAML MS unifies the fraud and AML/CFT layers, with no duplicate licenses or infrastructure.
go-live in 4 months
Integration with your core via API and Dynamic Message Mapping; visible value from the pilot run.
Your data stays with your institution
No transfer to third parties, in compliance with Law 1581 of 2012, Colombia's data protection law.
World-class infrastructure
Running on Oracle Cloud (OCI), with 99.9% committed uptime.
Ongoing support
Rule calibration, monthly continuous-improvement committees and 24×7 expert support for critical incidents.
Built for Latin America
Terminology and reporting aligned with the SFC (Colombia's Financial Superintendence), the UIAF (Colombia's financial intelligence unit) and regulators across the region.
From the first meeting to production, with a clear plan.
Discovery
Mapping of your institution's channels, risks and fraud typologies.
Configuration
Integration through Dynamic Message Mapping, segmentation, rules and models.
Testing and tuning
Validation in SandBox, training and knowledge transfer.
go-live
Production launch with ongoing support and continuous-improvement committees.
See FRAML MS AntiFraud running on your own fraud scenarios.
Book a 45-minute demo with our team of specialists. We'll walk you through the full flow: from the suspicious transaction to the decision.