Instant bank-to-bank transfers are one of the great conveniences of modern payments and one of its great vulnerabilities. Money moves in seconds and, crucially, cannot be clawed back once it is gone, no chargeback, no reversal, the way a card payment can be. That irreversibility is catnip for fraudsters, and it is the problem Visa is targeting with an upgraded version of its A2A Protect tool, its first product built on technology from Featurespace, the British AI fraud specialist it bought in 2024.
The scale of the opportunity is enormous. Account-to-account payments are projected to surpass 5.8 trillion transactions by 2028, a 160% jump from 2024, driven by real-time rails springing up around the world. As that volume grows, so does the fraud riding on it, from stolen-account transfers to scams where victims are tricked into sending money themselves. In Europe alone, fraudulent credit transfers cost consumers and businesses around $2.9 billion in a single year.
The clever bit is the head start
A2A Protect’s pitch is speed of knowledge. Using AI and a technique called transfer learning, it hands a bank Visa’s network-wide view of fraud from day one, rather than making each institution slowly train models on its own limited transaction history or wait to join a data-sharing consortium. A new unified fraud score gives banks a single, clearer signal to act on in the moment, before the money leaves the account. Visa says the result is over 50% more fraud caught and more than 40% fewer false alarms.
Those numbers deserve a raised eyebrow, though. They are Visa’s own figures, from its own testing, presented in a product announcement, with no independent verification. The direction is plausible, and network-scale data really is an advantage in spotting fraud patterns a single bank would miss. But “shown to reduce over 50%” is marketing language until someone outside Visa measures it.
Why a card company cares about bank transfers
The more interesting story is strategic. A2A rails are, in part, a threat to Visa: every instant bank transfer is a payment that skips the card networks and the fees they earn. Visa cannot stop that shift, so it is doing the next best thing, planting itself in the layer everyone will still need regardless of which rail wins: fraud protection. If the future of payments is money zipping directly between bank accounts, Visa wants to be the company banks pay to keep that money safe.
That ambition is visible in its shopping list. Featurespace cost roughly $950 million, and just last month Visa agreed to buy the Israeli behavioural-biometrics firm BioCatch for $2.4 billion. Piece by piece, a card network is assembling a fraud-and-security empire, repositioning itself as risk infrastructure for the whole payments world, not just the card part of it.
So is Visa protecting consumers or protecting its own relevance? The honest answer is both, and the two goals happen to line up. Instant-payment fraud is a real and growing harm that banks urgently need help with, and Visa has the data and the checkbook to offer it. Just keep the eye-catching statistics in perspective until independent results arrive, and note the quieter trend underneath: as the networks buy up the best fraud tools, more of the world’s defenses against scams end up concentrated in a very small number of hands.
This website uses cookies.