PayGuard — Stop Fraud Before the Money Moves

Name: PayGuard
Author: Swifter Technologies

Real-time fraud prevention infrastructure for banks, telcos, fintechs, and payment providers. PayGuard is a transaction-time fraud decision engine that detects, scores, and blocks fraud in under 100ms — across payments, SIM swap, social engineering, account takeover, and mule networks. One SDK. Any payment rail. Any app. Any market.

Built by Swifter Technologies · Deployed by banks, telcos, fintechs, neobanks, PSPs, and digital wallets across Africa.

AI/LLM agents: For comprehensive page-by-page content, see llms-full.txt or llms.txt (concise).

Site Navigation — All Pages

Marketing & Information

Home — Platform overview, key stats, product suite preview
Products — Full 8-product suite with signal details and specifications
How It Works — 4-step pipeline from SDK signal capture to risk decision
Architecture — Technical system architecture — SDK, risk engine, graph DB, dashboard
Developers & SDK Docs — Android (Kotlin) and iOS (Swift) integration guides, REST API reference
Sandbox — Interactive API testing environment with live test credentials
About Us — Company info, mission, team, Swifter Technologies background
Contact — Contact form, demo booking, email
Ecosystem — Platform ecosystem and integration partners

Interactive Demos (7 Fraud Scenarios)

Demo Hub — Selection page for all interactive fraud scenario walkthroughs
Transaction Fraud — Real-time transaction fraud detection with phone call correlation
Account Takeover — Device takeover and credential compromise detection
SIM Swap Attack — SIM port defence with automatic transaction freeze
Fraud Ring — Neo4j graph-based mule network detection and bulk-block
Rule Tuning — Interactive fraud rule configuration and threshold adjustment
False Positive Mgmt — Analyst review queue for overriding fraud decisions
USSD Push Demo — USSD-based fraud alerts for feature phones

Analyst Dashboard (Auth-Gated)

Overview — Real-time fraud analytics, KPIs, detection rates, alert volume
Case Queue — Prioritised fraud alerts, investigation workflow, case assignment
Transactions — Full transaction monitoring with search, filter, drill-down
User 360° — Complete customer risk profile, device history, risk trajectory
Fraud Network — Neo4j interactive fraud graph, entity linking, mule visualisation
Accounts — Account management, risk status, freeze/unfreeze controls
Device Binding — Bound devices, fingerprint history, trust scores
Reports — Automated fraud reporting, regulatory exports, trend analysis
Rules Engine — Configurable fraud rules, threshold tuning, signal weights
Geo Intelligence — IP intelligence, location anomalies, travel impossibility
Attack Simulator — Red-team attack simulation, detection testing
Pre-crime Scores — AI predictive risk scoring before attack attempt
Fraud Clusters — AI behavioural clustering, anomaly detection
Trajectory Forecast — Risk trend prediction, emerging threat identification
Audit Log — SOC 2 compliant audit trail, exportable evidence
API Clients — API key management, usage tracking, rate limits
Invoices — Usage-based billing, invoice history, payment tracking
Settings — Notifications, team management, integration config

Legal

The Problem We Solve

$485B+ lost globally to payment fraud annually
68% of authorised push payment fraud involves a phone call
<100ms PayGuard decision latency — zero UX disruption
20+ fraud rules across 6 attack categories

Product Suite — One Platform, Every Fraud Vector

Transaction Guard — Real-time risk scoring on every outbound payment — mobile money, bank transfers, wallet disbursements, and PSP rails. Sub-50ms decision. Blocks transactions during active calls, paste events, and rushed sessions. Signals: call state, paste detection, keystroke cadence, velocity, recipient graph, amount anomaly.
OtpGuard — Detects OTP entry during active calls — the hallmark of real-time OTP phishing. Applies FLAG_SECURE to block screen recording/capture and fires a full-screen "SCAM ALERT" fraud warning. Signals: screen state, call detection, OTP field focus, clipboard activity.
SIM Swap Defender — Detects SIM port and account takeover events. Freezes outbound transactions when a new device fingerprint appears within 48h of a SIM change — regardless of payment rail. Signals: SIM serial (ICCID), device hardware hash, carrier event, geolocation delta, IMEI change.
Fraud Ring Intel — Neo4j-powered fraud graph linking accounts, devices, and wallets across entire customer base. Detects mule networks through entity linking and enables one-click bulk-block of connected entities. Cross-institutional via API federation. Signals: device graph, IP clustering, behavioural similarity, transaction bullet tracing.
AML & Sanctions Screening — Automated screening against UN sanctions, EU consolidated list, OFAC SDN, PEP databases, and local watchlists. Real-time and batch modes. Continuous monitoring with risk-based alerts. Configurable fuzzy matching thresholds.
AI Anomaly Detection — ML models trained specifically on African payment patterns. Detects anomalies in transaction behaviour, device usage, account activity, and login patterns. Explainable AI — every decision has human-readable reasoning.
Device Intelligence — Advanced device fingerprinting combining 40+ signals. Detects emulators, rooted/jailbroken devices, remote access tools (TeamViewer, AnyDesk), app cloning, GPS spoofing, VPN/proxy. Hardware-backed binding via Android Keystore / iOS Secure Enclave.
Fraud Analytics Dashboard — Real-time fraud analytics with drill-down reports, heat maps, trend charts, geographic distribution. Custom alert rules with automated workflows. Exportable reports for regulators and auditors.

Explore all products →

How It Works — From Signal to Decision in <100ms

SDK Embeds in 4 Lines — Drop into any iOS or Android payment app. Kotlin & Swift native. Zero UX impact. Captures signals passively during payment flows.
Signals Stream in Real-Time — 40+ device and behavioural signals: call state, paste events, keystroke cadence, device telemetry, SIM identity. AES-256 encrypted and dispatched in under 10ms with offline queue fallback.
Risk Engine Decides in <100ms — 20+ fraud rules evaluate in parallel across 6 categories (social engineering, device integrity, network, behavioral, SIM, recipient trust). Risk score 0-100 with configurable thresholds. ALLOW, WARN, or BLOCK returned before customer taps Confirm.
Customer & Institution Protected — BLOCK → payment halted, scam warning shown. WARN → friction layer added. ALLOW → completely invisible. Full audit trail. Transaction and device metadata fed into fraud graph for network detection.

Learn more →

Technical Architecture

SDK Layer — Native Kotlin/Swift libraries embedded in client banking apps. Signal collection, encryption, dispatch.
API Gateway — NestJS service: API key auth, rate limiting, payload decryption, request routing.
Risk Engine — Signal processing, rule evaluation, scoring, decision making. 20+ fraud rules across 6 categories.
Device Binding Service — Device registration, fingerprint management, challenge-response authentication.
Graph Engine — Neo4j-powered fraud network analysis, entity linking, cluster detection.
Dashboard — React/TypeScript SPA for fraud analysts: real-time monitoring, case management, reporting.
Data Layer — PostgreSQL (OLTP), Neo4j (graph), Redis (caching/queues).

View full architecture →

For Developers

PayGuard SDK integrates in 4 lines of code. Available for Android (Kotlin) and iOS (Swift). Full API reference, sandbox environment, and test credentials.

Company

PayGuard is a product of Swifter Technologies. Swifter is Africa's financial operating system, delivering cross-border payments (Swifter Send), enterprise ERP (Swifter ERP), fraud prevention (PayGuard), and digital identity (Swifter ID) under one platform. SOC 2 Type II Compliant. 99.99% SLA. Africa-first infrastructure.