Skip to main content

Score every event in under 100ms

Device, behaviour, identity, and network signals fuse into one decision — PSD2 SCA-compliant, FCA APP responsibility scheme-ready. Chargebacks fall 40%; legitimate transactions pass without friction.

WIDTH fraud decisioning dashboard
<100ms
Scoring latency
Median latency from event ingest to block/allow action — within PCI DSS v4.0 real-time decisioning requirements.
7
Fraud types covered
ATO, synthetic identity, 1st-party fraud, mule networks, APP scams, CNP, and API abuse — each mapped to MAS Notice 655 or FCA APP responsibility scheme.
99.9%
Decision availability
Multi-region failover; fallback to signed policy defaults so PCI DSS v4.0 Req 10 audit trail is never broken.
40%
Chargeback reduction
Measured on first-30-day CNP volume across six production launches; issuer feedback loops updated daily per PCI DSS v4.0 Req 12.
7 typologies, one scoring engine

Channel-hop and the score follows you

One engine covers onboarding, login, payment, and withdrawal — attackers who pivot channels face the same risk score, not a fresh blind spot.

Account takeover

Detect ATO before the password changes

Device reputation, behavioural biometrics, and session-continuity signals fuse into a live ATO score. Mid-session step-up challenges trigger in 2 seconds — MAS TRM Guidelines §9.1 aligned — without blocking the 99.7% of honest sessions.

Explore ATO defence →
ATO defence
96%
ATO recall
0.3%
False challenge rate
2s
Step-up decision
Synthetic identity

Catch synthetic identities at bureau level

Cross-references identity fragments across credit bureau, device, and 30-day velocity signals — FATF Rec 8 non-face-to-face risk controls applied automatically.

See synthetic →
Scam + mule

Flag APP scams before funds leave

Beneficiary reputation, coercion-pattern signals, and peer-group velocity detect authorised-push-payment fraud — FCA APP responsibility scheme (Oct 2024) and MAS Notice 655 evidence packs generated automatically.

See scam defence →
CNP + chargeback

Pre-empt chargebacks before dispute filing

CNP scoring fused with issuer feedback loops updated daily — EU PSD2 SCA exemption logic applied at the transaction level to cut false declines alongside fraud.

See CNP →
API abuse

Block credential stuffing at the edge

Rate, pattern, and identity signals terminate credential-stuffing and scraping sessions in under 100ms — without rate-limiting legitimate API consumers or triggering PSD2 SCA friction.

See API defence →
Fraud rings

Surface organised fraud rings

Cluster accounts, devices, IPs, and beneficiaries in one graph — surfacing mule networks, romance-scam rings, and bust-out schemes that single-account scoring misses. Patterns scored against ACFE typologies feed straight into case management for parallel investigation.

See ring detection →
Static rules vs adaptive scoring

Fraud rings rotate tactics weekly

Models retrain daily on your labelled outcomes — every decision carries a one-click reconstruction trail for FCA APP and PCI DSS v4.0 dispute evidence.

Legacy rules
Rigid, reactive, noisy.
  • Rule changes lag attack evolution by 2–4 weeks
  • Over-broad CNP thresholds block 3–8% of genuine transactions
  • FCA APP evidence packs assembled manually per dispute
  • Onboarding, payment, and login each carry separate blind spots
WIDTH Fraud
Adaptive, targeted, explained.
  • Models retrain daily on labelled outcomes — attack lag under 24h
  • Peer-aware scoring keeps false-challenge rate at 0.3%
  • One-click PCI DSS v4.0 / FCA APP dispute reconstruction
  • Onboarding, AML, and fraud share one risk signal — no channel gaps

40% fewer chargebacks, 0.3% false challenges

30-minute session: we replay 7 days of your production events, score against 7 typologies, and quantify the chargeback reduction before you commit.