How fraud detection works
Each event is enriched with data and then judged, for a card payment in well under a second.
- Signals: device fingerprint, IP address, email and phone age, address match, card BIN
- Velocity: attempts per card, device or account in a period
- Rules: fixed conditions set by analysts, such as country mismatches or amount limits
- Models: machine learning trained on past fraud, giving a score
- Lists: known bad cards, devices and addresses, and trusted customers
- Decision: approve, decline, ask for 3-D Secure, or send to manual review
Where fraud detection sits
Several parties screen the same payment. The merchant or its gateway scores the order before authorisation, the issuer scores the authorisation request, and both look again afterwards for patterns across accounts. Banks, wallets and exchanges apply the same methods to logins, new accounts and outgoing transfers, where the threats are account takeover and scams.
Labels arrive late. A payment is confirmed as fraud only when the issuer reports it or a chargeback arrives, often weeks later, and those reports feed the next round of rules and training.
Why fraud detection matters to a payments product
Card schemes monitor the fraud reported against each merchant and acquirer, so weak detection ends in programme costs or lost acceptance, not only in direct losses. Where strong customer authentication applies, in the EU and the UK, detection also decides which payments are sent for authentication and which are put forward under an exemption.
Fraud detection, fraud prevention or transaction monitoring?
- Fraud prevention
- Controls that stop fraud from being attempted or completed: authentication, tokenisation, limits. Detection finds what gets through.
- Fraud scoring
- One technique inside detection: the number a model assigns to an event.
- Transaction monitoring
- The anti-money laundering duty to watch customers' activity for suspicious patterns. Similar tooling, a different legal purpose.
Where Fraud detection shows up in a platform
Modules that do the work
- Fraud & riskReal-time fraud scoring, limits and case management, with the risk engines you already trust.
- Analytics & monitoringProduct analytics, attribution, engagement and crash reports, on a tracking plan you control.
- AuthenticationSign-in, passkeys and second factors for customers and staff, on your identity provider or ours.
- CheckoutA hosted checkout for cards, wallets, bank payments and crypto, routed to the best provider.
Platforms that need it
Guides that go deeper
- How to become a PSP or a payment facilitatorGateway, acquirer, payment facilitator or payment institution: where the licence line falls, how the card schemes register payment facilitators, what merchant KYB demands, and the UK and UAE rules.
- Crypto on-ramps: how a buy widget works and who needs the licenceHow a fiat-to-crypto buy widget works inside another company's app, who the regulated party is, and the rules on UK promotions, card fraud, the Travel Rule, MiCA and VARA.