Mobile Money Feature Engineering Library
Out the box ML features across SIM, device, wallet, agent, account, and KYC signals.
- SOLUTIONS
- SERVICES
- ABOUT US
- RESOURCES
- CONTACT
FRAUD
MANAGEMENT
MANAGEMENT
BUSINESS
ASSURANCE
ASSURANCE
PARTNER
ECOSYSTEM
ECOSYSTEM
FRAUDZAPTM
Discover our legacy of advancing AI-led innovation and operational excellence across telecom and enterprises.
Meet the visionaries shaping
Subex’s strategy, growth, and future direction.
Explore how Subex is reshaping the future of telecom with purpose-driven, AI-led innovation.
Access insights, reports, and
updates for our valued investors and stakeholders.
Learn about our global network
of strategic partners driving transformation.
Join a team shaping the future of telecom intelligence through purposeful innovation and human-supervised AI.
Explore insights, ideas, and innovations shaping the digital
trust landscape.
Dive into in-depth research and
expert perspectives on emerging technologies.
Read thought leadership pieces
offering clarity on industry transformations.
See how Subex solutions drive
measurable impact across global enterprises.
Access quick overviews of
our offerings, capabilities, and value propositions.
Stay updated with the latest
fraud trends and prevention insights.
Catch up on Subex highlights,
innovations, and industry updates.
Join live and on-demand sessions
with experts driving digital transformation.
Stay updated with Subex’s latest announcements, product launches, and corporate milestones.
Explore how Subex is featured across global news, industry stories, and expert publications.
Overview
Subex Mobile Money FraudZap™ scores transactions in real time before funds leave the wallet, using wallet behaviour, agent risk, and SIM signals. Built for P2P transfers, cash-outs, merchant payments, and wallet credit, it helps telcos protect revenue, customer trust, and agent ecosystems. AI agents build cases, link suspicious patterns, draft STR narratives, and accelerate resolution.
The four ways mobile money gets attacked
Account Takeover
PIN reset followed by immediate high-value transfer before batch-based controls react
Wallet Credit Abuse
Borrow, cash out instantly, then default across mobile credit and wallet flows before controls fire
SIM Swap Cash-Outs
Funds move quickly after a SIM swap, often before customers notice. Batch tools fire too late
Agent Collusion
Agent-assisted transactions look legitimate unless behaviour, velocity, and relationships are modelled
The four ways mobile money gets attacked
Account Takeover
PIN reset followed by immediate high-value transfer before batch-based controls react
Wallet Credit Abuse
Borrow, cash out instantly, then default across mobile credit and wallet flows before controls fire
Agent Collusion
Agent-assisted transactions look legitimate unless behaviour, velocity, and relationships are modelled
SIM Swap Cash-Outs
Funds move quickly after a SIM swap, often before customers notice. Batch tools fire too late
Mobile Money FraudZap™ business benefits
Stop Fraud in real time
Score P2P, agent cash-out, merchant, and wallet-credit transactions before settlement so only accepted decisions proceed
Faster time to go live
Deploy across cloud, on-premises, or microservices. No cold start. No large data science team required
Accelerate Investigation with AI Agents
AI agents build fraud cases, link suspicious patterns, draft STR/SAR narratives, and close low-risk referrals automatically
High-accuracy risk scoring and detection
ML model scoring plus out the box rules detect SIM swap cash-outs, PIN reset abuse, agent velocity, account takeover, and wallet credit abuse from day one
Improve Decision Quality with prebuilt ML Features
FraudZap™ uses prebuilt ML features across SIM, device, wallet, agent, account, and KYC signals to strengthen real-time risk scoring
Support Risk, Finance, and Compliance Teams
Track savings, audit every decision, and report outcomes with decision logs, explanations, alerts, dashboards, and transaction-level audit trails
*Figures are indicative and may vary by customer based on business and operational conditions.
How it works
FraudZap architecture: transaction to decision in real time
Ten capabilities. One purpose: stop mobile money fraud
ML Model + Rules Hybrid Detection
Combines real-time ML scoring with pre-built rules for multi-layer fraud detection.
Real-Time Decisioning
Accept, decline, or refer suspicious transactions before settlement in real time.
Agentic Investigation
AI agents build cases, link suspicious patterns, close low-risk referrals, and draft STR/SAR narratives.
Full Wallet Journey Coverage
Supports P2P transfers, agent cash-outs, merchant payments, and wallet credit.
SIM and Device Risk Signals
Uses SIM swap recency, device changes, MSISDN age, and mismatch signals.
Agent Risk Monitoring
Tracks agent fraud rates, collusion, velocity anomalies, and off-hours activity.
AML and KYC Controls
Combines AML/CFT controls with KYC and biometric checks for stronger compliance.
Flexible Deployment Options
Runs on-premises, public cloud, private cloud, or microservices.
Explainable Decisions
Provides reason codes, alerts, dashboards, and transaction-level audit trails.
Choose your starting point
Start with real-time fraud scoring, add AML compliance controls as requirements grow, and advance to agentic investigation for automated case resolution.
Package 1
Real-Time Transaction
Fraud Scoring
- Synchronous ML scoring before settlement
- Pre-built mobile money ML features
- OOB rules: SIM swap, ATO, agent velocity, wallet credit
- Agent intelligence: collusion, off-hours, fraud rate signals
- SIM & device risk: swap recency, device change proximity
- Accept · Decline · Refer decisioning with reason codes
- No data science team required — pre-trained & ready
- Single API integration — no core system changes
Package 2
Package 1 + AML
Compliance Controls
Everything in Package 1, plus:
- AML/CFT compliance controls built into scoring
- Regulatory Mandate alignment
- KYC & biometric database integration for identity checks
- Explainable decisions: reason codes per transaction
- Full audit trail: every accept / refer / block logged
- Dashboards for risk, finance & compliance teams
- Supports regulatory reporting obligations
Package 3
FraudZap Agentic
From fraud flag to closed case – automatically
- AI agent auto-builds the fraud case
- Agent ring detection and pattern linking
- STR narrative drafted in seconds
- Low-risk referrals closed automatically
- Natural language investigation interface
Frequently Asked Questions
FraudZap sits alongside your existing platform as a real-time scoring layer. It connects via a single API call and requires no changes to your core mobile money system or existing FMS.
Yes. Every transaction is scored synchronously before settlement in under 300ms. There is no batch window, no post-event review step, and no delay in the payment flow from the end user's perspective.
Package 2 includes AML/CFT compliance controls aligned to central bank mandates in UAE, Jordan, Egypt, and Saudi Arabia, with full audit trails, KYC integration, and regulatory reporting support built in.
FraudZap uses 60+ features across SIM events, device signals, wallet transaction history, agent behaviour, account profile, and KYC data. It does not require external data enrichment vendors to operate effectively.
FraudZap runs on public cloud, private cloud, on-premises infrastructure, or as microservices depending on your requirements. Deployment takes 3–4 weeks with no large data science team required.
Yes. Every decision (accept, refer, or block) is logged with the full decision context: risk score, signal values, model version, reason code, and timestamp. Package 2 adds compliance dashboards and regulatory reporting exports.
FraudZap™ Agentic builds the fraud case, links agent rings and suspicious patterns, drafts STR/SAR narratives in seconds, and can automatically close low-risk referrals. Investigators can also use a natural-language interface to support investigation.
