Ten seconds. That’s the new legal threshold set by the European Union’s Instant Payments Regulation¹. Once fully in effect, the law will require banks and payment providers to make instant transfers available to recipients within 10 seconds – 24/7, across borders and at no higher cost than a standard transfer. It also introduces pre-authorization checks, meaning fraud detection must work in real time – within just 50 milliseconds. In the blink of an eye, the system must spot anomalies, interpret risk signals, decide whether to block or approve the transaction, and do it all without slowing the customer experience.

The economic toll of cybercrime is staggering. In Germany alone, losses reached an estimated €267 billion (US$310 billion) in 2024². Across Europe, studies show that fraud attempts involving instant payments are climbing sharply – often doubling year over year. Regulators are taking notice. Germany’s financial supervisory authority, BaFin, has urged banks and payment providers to use every tool at their disposal to detect and block fraudulent transactions before money leaves an account.

For Deutsche Kreditbank (DKB), it was a clear signal to rethink how fraud prevention should work in a world of instant payments. As Germany’s second-largest direct bank, DKB serves nearly 6 million customers and processes up to half a million transactions every day. That scale brings both opportunity and exposure. To stay ahead of fraudsters moving just as quickly as instant payments themselves, the bank decided to modernize its system from the ground up.

Building on its long-standing collaboration with SAS, DKB transformed its fraud detection system into a cloud-native, AI-driven platform. Using SAS Financial Crimes Analytics on SAS Viya deployed with SAS Managed Cloud Services on AWS, the bank now combines real-time decisioning, machine learning and integrated alert and case management – allowing it to analyze and score each transaction in just 50 milliseconds. That speed is critical to meeting the EU’s 10-second requirement while keeping legitimate transfers seamless. By combining advanced analytics with a decision-first architecture, DKB can catch more fraudulent activity, reduce false positives and minimize customer friction.

¹Source: https://www.ecb.europa.eu/paym/retail/instant_payments/html/instant_payments_regulation.en.html

²Source: https://www.biocatch.com/report/digital-banking-fraud-trends-germany-2025-de

Modernizing for milliseconds – inside DKB’s cloud transformation

The shift from DKB’s on-premises system to a cloud-based platform took about a year. “It was a focused, steady transformation,” says Florian Lindemann, Head of Fraud Analytics and Data Platform at DKB. “It began with two key decisions. First, we made a company-wide commitment to the cloud – it’s a fundamental part of our strategy. Second, we chose to modernize to SAS Viya as a managed service in the cloud.” The goal: to stay ahead of increasingly sophisticated fraud attempts while lowering operating costs.

The bank’s new fraud detection system is deployed by SAS Managed Cloud Services on AWS, which fully supports SAS Viya. SAS Business Orchestration Services handles the data orchestration in real time, pulling in the live signals that matter (e.g., account behavior, device/session data, payee/name-IBAN checks, lists/limit), normalizing them and then making them available to SAS Viya for real-time decisioning. It also handles the API calls back and forth with the core banking platform.

In SAS Intelligent Decisioning, fraud specialists develop and refine automated decision flows, and SAS Model Manager enables data scientists to build, deploy and continuously improve machine learning models for anomaly detection. “Our teams aren’t limited to coding in SAS,” adds Lindemann. “They can seamlessly use open source languages like Python as well.”

AI and machine learning were essential requirements for the new infrastructure. “Without anomaly and pattern detection, and without the learning capabilities of AI, fraud detection in the age of instant payments would be impossible,” says Lindemann. “SAS Viya on AWS delivers fraud insights at scale, and DKB can enrich models with trusted third-party data to materially increase detection accuracy.”

And when a transaction needs human follow-up, SAS Visual Investigator keeps casework moving – clear alert triage, status at a glance and faster investigations – so customers get immediate notification of attempted fraud and quick, confident resolution.

Protecting customers without slowing them down

Accuracy is crucial for one of fraud detection’s core requirements: minimizing false positives.

“Imagine that one of our customers wants to buy a used car and pay the seller via instant transfer,” Lindemann says. “If the payment is blocked despite being legitimate, the delay can derail the deal and cause real frustration.”

That balance between security and convenience is what drives DKB’s approach.

“Precision in fraud detection is essential,” Lindemann concludes. “With the SAS Viya platform, we can safeguard legitimate customers and identify real fraudsters, detecting suspicious activity as early as the registration stage.”