Visa has launched an updated A2A Protect solution, integrating Featurespace technology to detect fraud before funds leave accounts.
The expanded version includes a unified fraud score, marking the integration of technology from Featurespace, the UK-based fraud analytics company that Visa acquired in 2024.
Rising A2A payment volumes drive fraud prevention focus
A2A payments, which move funds directly between bank accounts without card networks, are expanding rapidly worldwide. According to the data cited by Visa, A2A transaction volumes are projected to reach USD 5.8 trillion by 2028, representing a 160% increase from 2024 levels. This growth has increased pressure on financial institutions to detect fraud in real time, as A2A transfers are typically irrevocable once completed, unlike card payments that carry chargeback mechanisms.
According to the official press release, A2A Protect applies artificial intelligence and transfer learning techniques to provide banks with risk insights drawn from data across Visa's network, instead of relying solely on a single institution's historical transaction data. Visa states this allows banks to access risk detection capabilities without the lead time typically required to train models on proprietary data or the need to wait for other institutions to join a data-sharing consortium.
Network-level intelligence and integration
Financial institutions that opt into network-level intelligence sharing can access additional signals designed to flag emerging scam patterns and coordinated fraud activity across the payments ecosystem. With this in mind, Visa states this shared visibility can help identify scams before transactions are authorised, addressing activity that may be harder for individual institutions to detect using their own data alone.
Additionally, Visa reports that A2A Protect increased fraud detection by 75% within the first six months of deployment. The company also adds that the tool has been associated with a reduction of over 50% in fraud and over 40% in unnecessary fraud alerts.
The solution connects to banks' existing systems through a single API, which Visa says is intended to reduce implementation time and complexity compared with building bespoke detection infrastructure. Each fraud alert generated by the system includes a plain-language explanation of the flagged risk factors, aimed to help fraud teams assess cases more efficiently without introducing friction for legitimate customers.
The rollout reflects a broader industry shift toward real-time fraud prevention infrastructure as A2A payment methods, including Open Banking-enabled transfers and instant payment schemes, continue to gain adoption globally. For example, Navan and DataVisor partnered recently in 2026 to offer a scam-free experience for Navan’s customers, by using an adaptative machine learning combined with rules that detect fraud patterns.
