AI to reshape bank payments as fraud pressure mounts

Financial technology firm RedCompass Labs expects banks to push artificial intelligence deeper into day-to-day payment operations in 2026, while regulators and platforms face mounting pressure over the rise of organised fraud.

The company's senior executives forecast a sharp shift from experimentation to large-scale deployment of AI in payments, as institutions redesign ageing systems and respond to rising scam losses. They also predict a widening performance gap between early adopters and slower-moving rivals.

AI in operations

Oliver St Clair-Stannard, VP, Payments AI Strategy & GTM at RedCompass Labs, said banks had moved beyond pilots and proofs of concept during the past year.

"The pace of change in payments this year has been unprecedented, and 2026 will move even faster. Over the past 12 months, we've seen AI shift from experimental to operational. In 2026, payment teams will routinely use AI models trained on institutional knowledge and bank documentation that, until recently, was inaccessible to large language models.

St Clair-Stannard said internal AI assistants would take on a growing share of documentation work.

"AI assistants, powered by multi-agent systems engineered for accuracy over speed, will prepare, review and deliver modernisation documentation in a fraction of the current time. Analysts will spend less time typing and more time designing and validating. Those that deploy Applied AI tuned specifically for payments and guided by domain experts will cut costs and project timelines in half.

He said early adopters were already seeing a change in programme delivery.

"Banks that embraced AI in 2025 are already accelerating transformation programmes and moving ahead of competitors. Those that hesitated will now face higher costs, slower change and growing pressure from faster-moving rivals, creating a large gap between first movers and laggards.

St Clair-Stannard said institutions could still benefit if they focused investment and oversight.

"If banks harness the billions already invested into AI technologies, apply sector-specific training and combine it with expert oversight, they have a genuine chance of keeping pace. Next year, the winners will be using AI intelligently, for the right problems, with the right expertise behind it."

Modernisation pressures

Payments systems worldwide are in the middle of a long-running migration to real-time infrastructure and data-rich ISO 20022 messaging standards.

Pratiksha Pathak, SVP, Head of Payments - UK at RedCompass Labs, said the past year marked a turning point in this process.

"2025 marked the moment the payments industry shifted from experimentation to execution. SEPA Instant gained real traction, the Eurosystem's plan to link TIPS with India's UPI signalled a new phase of global interoperability, and regulators across the US, EU and Canada took concrete action to bring stablecoins into the mainstream. The year concluded with a landmark event: the end of MT-MX coexistence. After two decades, cross-border payments are finally being required to speak a modern, structured, data-rich language.

Pathak said much of that work remained tactical.

"But it's important to recognise that much of this year's progress has been tactical-a race to hit deadlines, apply quick fixes and keep legacy systems compliant. That phase is ending. Organisations are now stepping back and asking harder questions: can technologies built 40 years ago really support the next generation of real-time, intelligent payments? Increasingly, the answer is no.

She said institutions would now reassess their underlying architectures.

"In 2026, payments modernisation will remain at the top of the agenda, but the conversation will change. The industry is moving from incremental upgrades to fundamental redesign. AI will be the catalyst. It will enable sharper fraud detection, more adaptive compliance, and faster, data-driven decision-making-capabilities that simply aren't possible on legacy architectures.

Pathak said direct integration of AI into payment flows would become a key differentiator.

"Banks that integrate AI-powered insights directly into their payment flows will separate themselves from the pack. They won't just meet regulatory expectations-they'll improve margins, reduce operational drag, and create more compelling customer experiences. Those that do not move quickly risk falling behind as payments becomes a real-time, intelligence-driven utility.

She said institutions that continue to treat payments as a back-office function would be exposed.

"2026 will reward the institutions that treat payments as a strategic differentiator, not a cost centre. The competitive gap between modernised and non-modernised banks is about to become very visible."

Fraud moves upstream

Fraud losses have drawn sustained attention from UK regulators and industry bodies, with authorised push payment scams and online investment schemes among the fastest-growing threats.

Silvija Krupena, Director of Financial Intelligence Unit at RedCompass Labs, said the recent trend had been negative despite earlier momentum.

"Fraud prevention slipped backwards in 2025, just as criminals have industrialised their operations. The momentum of 2024, driven by rising losses, fast-evolving scam models and growing recognition of the threat, has not carried into 2025. The delay to the National Fraud Strategy leaves millions exposed, and signs that platform accountability may be softened are deeply concerning. Fraud begins online. Any strategy that avoids that reality is not a strategy.

She pointed to recent industry loss figures and the impact on victims.

"UK Finance reported £629.3 million stolen in the first half of this year, an increase on 2024. Behind that figure are people whose savings, relationships, mental health and sometimes lives were taken by organised criminals operating with near-total freedom across social media, messaging apps and telecom networks. Banks are fighting back, but the fastest-growing scams, like investment, romance and impersonation fraud, are designed to bypass the banking system entirely.

Krupena said banks could not carry the burden alone and called for earlier intervention by digital platforms and telecom operators.

"In 2026, we must shift the fight upstream. Tech platforms and telecoms need to stop fraudulent ads, cloned sites and fake accounts before they reach the public. Whistleblower claims that a significant portion of Meta's ad revenue is tied to illegal activity show the scale of the problem. When companies profit from the same digital spaces where victims are targeted, accountability cannot be optional. The EU's decision to require platforms to compensate victims is an important milestone.

She said fraud now sat alongside other major organised crime revenues and required a broader policy response.

"Fraud is now one of the UK's largest criminal revenue streams and the most common crime affecting the public. Treating it as a payments issue has never made sense. The priority must be a whole-ecosystem response. We need upstream responsibility, stronger digital safety standards, better data sharing and major public awareness efforts. To protect the public, we must disrupt the system that enables fraud to spread so quickly in the first place."