Fraud prevention in Germany

Fraud prevention in banking is the work of stopping payments that the account holder did not want, or wanted only because someone deceived them. Two patterns dominate. In the first, the fraudster issues the payment: with stolen card data online, a stolen card in a shop or hijacked online banking access. In the second, the customer sends the money personally after a fake bank call, a false invoice or an investment scam. The joint EBA and ECB report on payment fraud puts reported payment fraud in the EEA at EUR 4.2 billion for its latest full year, and finds that manipulation of the payer accounts for more than half of the value of fraudulent credit transfers.

That split decides who pays. Under section 675v of the German Civil Code, a customer bears at most EUR 50 of an unauthorized payment with a lost or stolen payment instrument, unless the customer acted with intent or gross negligence, and nothing at all if the bank did not require strong customer authentication. A payment the customer authorized personally, even under deception, is harder to recover. The EBA and ECB report shows the gap across the EEA: payment service users bore around 85 percent of fraud losses on credit transfers, against 38 percent on card payments.

Two EU rules shape the daily work of fraud teams at German banks. Strong customer authentication under the Payment Services Directive (PSD2) asks for two independent factors, such as a phone and a PIN, before a payment. According to the EBA and ECB report, card payments with SCA show lower fraud rates than those without, and card fraud rates were about seventeen times higher when the other side of the payment sat outside the EEA, where SCA may not apply.

The second is verification of payee, required by the EU instant payments regulation. Before a customer authorizes a euro transfer, the bank compares the payee name with the IBAN and reports a match, a close match or no match. BaFin writes that the check aims to reduce misdirected transfers and to prevent fraud. If the customer goes ahead after a no-match warning and the money reaches the wrong person, the bank is not liable; if the bank skipped the check, it must refund. The instant payments answer covers the dates and the daily sanctions check that the same regulation adds.

Fraud detection at banks runs on transaction monitoring: a model scores each payment against the customer's usual behavior, the device, the location and the payee, and holds the risky ones for review. Machine learning models do much of this scoring next to fixed rules. Several German firms build this software. Hawk, founded in Munich, uses AI to detect transaction fraud, scams and money laundering and aims to cut unnecessary alerts. Risk Ident in Hamburg, a subsidiary of the Otto Group, detects fraud patterns in online transactions for e-commerce, telecoms and financial services.

Fraud prevention and anti-money laundering sit close together. A scam has a victim on one side and, on the other, often a money mule account that receives the stolen funds, which is why a vendor such as Hawk sells fraud and AML detection on one platform. Identity checks at onboarding, covered on the KYC in Germany page, are the first line against accounts opened with stolen or synthetic identities.

Not all fraud comes from outside. Wirecard, a German payment processor, collapsed after it disclosed that EUR 1.9 billion was "missing" and filed for insolvency. Nearly half of its transaction volume had run through opaque third-party acquirers, BaFin lacked the authority to examine its core business, and the auditor failed to verify the cash. For compliance teams in Germany, the case is the standard example of why fraud risk covers accounting, partners and auditors as well as customer payments.

Finance Loop is the meeting place for fraud analysts, payment product teams and the data scientists who train detection models. It connects the finance, IT and AI communities, with events in Frankfurt and from time to time in Munich, Berlin and Hamburg.

Fraud detection was one of the practical topics at KI Exchange in Hamburg, a conference by Payment & Banking on AI in finance, next to AI governance and DORA compliance. Finance Loop supports When Banks Say 'No', a Frankfurt seminar where Albert Quehenberger of A|Q FORENSICS shows how blockchain forensics traces crypto flows in fraud cases. Related pages: sanctions compliance in Germany and MiCA in Germany.

Fraud prevention in banking covers everything a bank does to stop payments the customer did not intend: strong authentication at login and payment, device and behavior checks, transaction monitoring with rules and models, verification of payee, limits for new payees, warnings in the app and a team that calls customers before a suspicious transfer leaves. It also covers recovery, when the bank tries to recall money from the receiving bank.

APP fraud stands for authorized push payment fraud: the customer is tricked into sending a transfer personally, for example by a caller posing as the bank's security team or by a fake investment site. The payment is authorized, so the cap for unauthorized payments in section 675v of the German Civil Code does not apply. Verification of payee and in-app warnings target this pattern.

If a card or online banking access is at risk, call the bank or the Sperr-Notruf 116 116, which blocks the girocard, credit cards, online banking access and the electronic ID card around the clock. Then file a report with the police, online or at a station, and keep the payment details and messages for the bank.

Fraud analyst and fraud prevention jobs in Germany are at banks, card issuers, payment institutions, online retailers and the software firms that build detection tools, in cities such as Frankfurt, Munich, Berlin and Hamburg. The work combines investigation, data analysis and customer contact. Data scientists who tune detection models increasingly sit in the same teams.