Agentic Fraud Is Shrinking The Defence Window, Yet Most Institutions Are Behind

We’re living in a time when convenience is moving as quickly as you can blink, including all things payments, thanks to the gift of AI. Yet the very same technology is fuelling something far more sinister and constantly proving harder to tackle: fraud.

Fraud numbers are growing exponentially. According to Juniper Research, eCommerce fraud could rise from $56 billion in 2025 to hit highs of $131 billion in 2030, with factors like friendly fraud carving deeply into merchant margins.

Visa paints an equally alarming picture. Today’s cyber criminals are arriving armed with AI-powered deepfakes, synthetic identities, and agentic scams that can hijack a customer’s entire identity through hyper-realistic impersonations.

Once in, they own every transaction that follows. This is fraud at scale, and the world’s fraud teams are operating in right now.

The scariest fact is that most of the systems sitting between customers and their money were not quite built for this challenge, and needed to catch up yesterday.

The New Fraud Frontier Is Agentic

If 2024 was the year of AI-generated deepfakes, 2025 is the year fraud went pretty much fully autonomous with AI.

Agentic AI, software that can reason, plan and execute decisions across multiple steps without human intervention, has become the payment industry’s most exciting and dangerous pet.

We’re seeing solutions like Mastercard’s Agent Pay surface, which is ‘enabling secure, scalable and trusted payments in agentic commerce’.

AI-initiated payment solutions are becoming a structural feature in modern commerce. The same technology enabling that convenience is right now being exploited by bad actors, running fraud operations at a speed and scale human teams are finding harder to match.

Think about it for a moment. When a fraudulent agent enters the picture, it does not merely mimic a person. It mimics a trusted AI agent, introducing an entirely new layer of deception that most detection engines may not be calibrated to catch.

Synthetic identities, which are trained to exhibit good customer behaviour across months and even years, are among the hardest to detect, simply because they do not look fraudulent until the moment exploitation hits.

For organisations, the implications are acute. As real-time payment rails expand across the region, driven by national switches and the rapid growth of instant and A2A payments, the attack surface expands with them.

Faster payments are an unambiguous benefit for consumers, and equally beneficial for fraudsters. Your institution is possibly already facing some facets of AI-orchestrated fraud. The next question is whether your fraud stack is built to meet it or built for yesterday’s threats.

SmartVista Fraud Management, Architected for the Moment

The BPC SmartVista Fraud Management was built for this environment. Trusted by more than 200 clients across 80+ countries, SmartVista Fraud Management is an enterprise-grade, omni-channel AI fraud management platform that unifies real-time scoring, machine learning, behavioural profiling and link analysis into a continuous flow.

SmartVista Fraud Management delivers a 360-degree view of the customer across all channels, regardless of whether a customer uses their card online, in-store, in-app, and in every jurisdiction.

Underpinning this is a continuous profiling engine: every piece of information is monitored, used and reused, creating a precise customer profile built from common behavioural patterns.

For enterprises that can’t afford a rip-and-replace approach, BPC provides a dedicated integration layer supporting any interface or API. SmartVista Fraud Management works alongside what you already have, saving both cost and time without disrupting existing middleware.

Its architecture is also designed for the pace at which modern fraud operates. Built on microservices and deployable on Kubernetes across AWS, Google Cloud, OCI, or on-premise, SmartVista scales elastically with transaction volumes.

Detection rules can be updated in minutes, critical when fraud patterns mutate faster than traditional release cycles allow. Maxim Kuzin, Head of Fraud Prevention and Risk for BPC Banking Technologies, shares,

“Our goal isn’t just catching more fraud, it’s giving the future-proven tools to our customers to prevent the most modern fraudsters and shorten the time to contain them. Update your defences in minutes, not days or months, and stay aligned with local compliance wherever your business is operating.”

In instant payment rails, the window between a fraudulent transaction and an irreversible transfer is seconds, and that window is unforgiving.

SmartVista’s online, real-time risk and fraud prevention solution is built to close it, a capability that is central to the platform since its launch in 2008 and still remains its defining priority today.

Proven in the Field With Malaysia’s Co-opbank Pertama

When Bank Negara Malaysia issued stricter directives mandating all licensed banks to adopt high security benchmarks for digital channels, Co-opbank Pertama (CBP) needed a platform that could meet requirements like velocity checks to analyse customer behaviour without disrupting operations or customer experience.

CBP deployed SmartVista Fraud Management across its retail and corporate mobile and internet banking channels.

The platform delivers real-time transaction monitoring alongside near-real-time and offline validation modes, behavioural profiling, rules-based analysis, comprehensive case management, and a notification engine, all surfaced through a 360-degree view of data via customisable dashboards and reports.

Zairil Anuar Ahmad, Chief Technology Officer at CBP, commented,

“Implementing high-standard preventive measures is both a top priority and a regulatory requirement as Malaysia intensifies efforts to combat fraud. Through our partnership with BPC and the integration of their advanced fraud management solutions, we have strengthened our defence mechanisms and are now better equipped to deliver a secure, seamless digital banking experience.”

A machine learning module was also introduced to leverage historical fraud data and continuously refine detection mechanisms, sharpening the precision of the bank’s defences. The entire project went live in under four months.

CBP’s experience reflects a pattern playing out across the region: regulatory requirements and fraud threats are not separate problems requiring separate responses.

They are the same problem, and financial institutions that address them through a single adaptive platform are the ones that stay ahead of both.

Strengthen Your Fraud Defences With SmartVista Fraud Management

Fraud teams do not need another reminder that the threat landscape has evolved. What they need is the ability to respond at the same speed and scale, if not more than the attacks that they’re facing.

SmartVista Fraud Management was built for this shift, enabling institutions to move from fragmented detection to real-time, omnichannel protection without the operational drag of legacy systems. Whether the challenge is rising alert volumes, increasing false positives, or the latest fraud AI agent, the solution is designed to address it effectively.

BPC works with tier-1 banks, national switches, payment service providers, fintechs, and government agencies, and the breadth is intentional, as fraud is a global problem. The intelligence that comes from operating at that scale, across regulatory environments and channel types, is embedded into how the platform develops.

Assess your fraud readiness and see how BPC’s SmartVista Fraud Management can deliver real-time, adaptive protection across every channel.

Featured image edited by Fintech News Malaysia based on images by ezps and stocksbuddy on Freepik