Fraud and financial crime
Fraud detection for fintechs and digital wallets: mule accounts, account takeover and promo abuse
Fintechs and digital wallets detect fraud by scoring every onboarding, login and transfer in real time, using device fingerprints, behaviour and the network of who sends money to whom. The main use cases are mule accounts, account takeover, social engineering scams, synthetic identity onboarding, merchant fraud and promotion abuse. Graph features catch rings that single-transaction rules miss.
Published . Last updated
On this page
What fraud do fintechs and wallets face?
Wallets move money instantly and onboard customers in minutes. That speed attracts fraud.
- Mule accounts: wallets opened or rented to receive and move stolen funds.
- Account takeover: stolen credentials or one-time codes, often after SIM swap or phishing.
- Social engineering: customers persuaded to send money to scammers.
- Synthetic identity onboarding: accounts opened with mixed real and fake details.
- Merchant and agent fraud: fake merchants, refund abuse, agent cash-out schemes.
- Promotion abuse: many accounts created to claim cashback or referral bonuses.
Why does network analysis matter?
Fraudsters work in rings. One device opens ten wallets. Ten wallets send money to one account that cashes out. Each transfer looks normal on its own. A graph of devices, phone numbers, accounts and counterparties shows the ring.
Useful graph features include:
- Number of accounts that share a device or phone number
- How many unrelated senders paid an account in the last 24 hours
- How fast money leaves an account after it arrives
- Distance in the graph to an account already confirmed as fraud
How fast does scoring need to be?
Transfers need a decision in milliseconds to seconds. Onboarding can take a little longer. Mule detection can run every few minutes across the network. On DataNature:
- StreamHouse keeps rolling features per account, device and counterparty.
- Lakebase serves the latest features to the scoring API.
- DataNature AI scores each event and returns a decision.
- Graph features are recomputed on the lakehouse and pushed back into Lakebase.
How do you handle promotion abuse?
Link sign-ups by device, payment method, address and referral chain. Score clusters, not single accounts. Hold the reward, not the account, when risk is high, so real customers are not punished.
What about regulation?
Wallet and payment operators report suspicious activity to their financial intelligence unit and keep records for regulators. Fraud and AML teams should work from the same data. See FRAML: why fraud and AML teams should share one platform.
Related
How does StreamHouse power wallet fraud detection in real time?
Wallet fraud moves at the speed of the app, so detection has to as well. StreamHouse runs the hot path:
- Events from the wallet backend (logins, device registrations, top-ups, P2P transfers, merchant payments) and change data capture from the ledger database stream into managed Kafka, with event contracts in the schema registry.
- In-flight features keyed by wallet and device: transfers per minute, new beneficiaries, device sharing across accounts, rapid top-up then cash-out.
- Real-time scoring with a registered model, plus rules for known mule patterns.
- Actions in seconds: block a transfer, require step-up authentication, or freeze a device across all wallets linked to it.
- Write-back of every event and decision to Iceberg for audit, partner-bank evidence and retraining.
Batch still matters: nightly Connect loads and dbt models build the slower features (account age, historical behaviour, network links between wallets) that the hot path reads at scoring time.
How does DataNature Atlas keep this secure and compliant?
Investigators see what they need; analysts see masked identifiers; every score and decision is auditable. DataNature Atlas is the platform's governance module:
- Knows your data: a catalog with automated PII detection and classification across Bronze, Silver and Gold.
- Controls access: role-based grants plus attribute-based policies, row filters and column masks (null, redact, partial, hash, year only).
- Governs AI: Voir agents are scoped principals; masks, row filters and PII guardrails apply to every tool call and answer.
- Proves it: lineage from source to metric to decision, and an audit log of who or what accessed which data.
- Stays local: Atlas runs inside your deployment, so policies, lineage and logs never leave your jurisdiction.
These controls support GDPR, PIPEDA, Quebec Law 25, CCPA, HIPAA and sector rules such as OSFI E-23. Read DataNature Atlas in depth or see security and compliance.
Platform components in this article
Learn more about each DataNature product and module mentioned here.
See it on your own data.
Book a demoDemos and quotes are handled by Dlytica, the company behind DataNature.