Financial institutions are facing a fraud crisis that traditional detection systems simply cannot keep pace with. Criminals are now using artificial intelligence to launch attacks at speeds and sophistication levels that rule-based fraud systems struggle to catch, leaving banks vulnerable to increasingly complex schemes like authorized push payment (APP) fraud and AI-generated social engineering attacks. Experian has launched a new solution called Transaction Forensics that combines advanced AI models with comprehensive identity data to detect fraud in real time, addressing what the industry calls the "detection gap".

What Is the "Detection Gap" in Modern Banking Fraud?

Banks face three interconnected problems when trying to stop financial crime. First, sophisticated fraud attacks, including mule networks and AI-generated social engineering, are evading detection at alarming rates. Second, when banks tighten security to catch more fraud, they inadvertently flag legitimate transactions, frustrating genuine customers and forcing teams to manually review thousands of harmless payments. Third, regulators now require that any AI used for fraud detection must be explainable, meaning banks cannot simply deploy a "black box" system and hope it works.

Transaction Forensics addresses these challenges by combining more than 80 AI models with Experian's proprietary consumer and commercial data assets. The system enriches transaction signals with insights across credit, identity, fraud, and anti-money laundering data, alongside historical behavioral patterns. This allows financial institutions to assess risk with precision at the transaction, customer, or company level, rather than relying on rigid rules that miss new attack patterns.

How Does the New AI Fraud Detection System Actually Work?

• Real-Time Analysis: Transaction Forensics analyzes bank-to-bank payments as they happen, using behavioral analytics to understand the true intent behind each transaction rather than just flagging suspicious keywords or amounts.
• Multi-Layer Data Integration: The system combines Experian's identity and credit data with transaction patterns and anti-money laundering signals, creating a comprehensive view of fraud risk that traditional systems cannot match.
• Support for Multiple Payment Types: The solution works across Faster Payments, BACS, and CHAPS payment systems, enabling detection of APP fraud, money mule activity, and money laundering schemes at scale.

The system is designed to complement existing fraud prevention tools rather than replace them entirely. Banks can deploy it as an additional analytical layer within their current systems or apply it selectively to high-risk transactions, giving institutions flexibility in how they integrate the technology.

What Results Has the System Achieved in Testing?

Pilot testing of Transaction Forensics demonstrated significant performance improvements across three critical metrics. The system achieved a 200% increase in APP fraud detection, meaning it caught twice as many fraudulent authorized push payments compared to traditional systems. Equally important, it reduced false positives by 80%, dramatically cutting the number of legitimate transactions flagged as suspicious. This also led to a 50% reduction in total alert volumes, allowing investigation teams to focus their limited resources on genuine threats rather than spending time on false alarms.

"Financial services are facing a significant challenge in identifying and stopping fraud and financial crime attacks, which are increasingly enabled by AI and at a scale not seen before. Transaction Forensics harnesses the power of AI to help businesses meet that challenge head on," said Paul Weathersby, Chief Product Officer at Experian UK&I.

Paul Weathersby, Chief Product Officer, Experian UK&I

These improvements matter because they address the core tension in fraud prevention: catching more fraud without creating friction for legitimate customers. The 80% reduction in false positives is particularly significant, as it means fewer genuine customers will experience delayed payments or account freezes due to overzealous security measures.

Why Is Explainability Important for AI Fraud Detection?

The Financial Conduct Authority (FCA), the UK's primary financial regulator, has issued guidance requiring that AI systems used for fraud detection must be explainable. This means banks cannot simply deploy an AI model and tell customers "the algorithm said no." Instead, institutions must be able to explain why a transaction was flagged or approved, supporting what regulators call "Consumer Duty" outcomes. Transaction Forensics is built with this requirement in mind, making it easier for banks to meet both regulatory expectations and customer trust standards.

"The use of AI in fraud and financial crime prevention is no longer optional but essential. By combining Resistant AI's advanced models with Experian's leading datasets, we are enabling financial institutions not just to address current attacks including APP fraud and money laundering but any new threats which will undoubtedly emerge in the years ahead," stated Martin Rehak, Chief Executive Officer at Resistant AI.

Martin Rehak, Chief Executive Officer, Resistant AI

Transaction Forensics launched in April 2026 and is now available to UK financial services institutions. The solution represents a partnership between Experian and Resistant AI, a fraud detection specialist that Experian invested in during July 2025. Additional products from this collaboration are expected to launch throughout 2026, suggesting that AI-powered fraud detection will continue to evolve and expand across the financial services industry.