XML-Based Messaging Tech Extends Fraud Detection Into Wider Bank Use Cases

Authorized push payment, or APP, scams are among the hardest frauds to stop. APP scams rely on the victim's voluntary authorization of the payment. Scammers employ impersonation, social engineering and artificial intelligence to make the outreach look genuine. Since the victim initiates the transaction, it is often difficult or impossible to reverse the payment once made. Transactions are completed in a few seconds, reducing the window of opportunity to spot and stop scams.

See Also: Ping Identity: Trust Every Digital Moment

Part 1 of this two-part series explored the rising threat of APP scams in real-time payments, or RTPs, and the role of network intelligence in containing them. Part 2 focuses on how ACI's "Signals" are being deployed to combat fraud in practice.

As criminals exploit impersonation, social engineering and AI-driven deception, ACI Worldwide is betting that its Signals network intelligence can turn the tide, detecting fraud in real time without requiring banks to share data.

An ACI study titled "2024 Scamscope: The Battle for Trust" showed that APP fraud through RTPs is predicted to increase from 63% of all APP fraud losses in 2023-2024 to 80% by 2028.

Financial institutions have made huge investments in multifactor authentication, which has largely curbed financial fraud. But these measures fall short in combating APP scams.

Cleber Martins, head of payments intelligence and risk solutions at ACI Worldwide, said defeating such scams demands more than stronger controls. "Beating scammers requires shared responsibility and systems that eliminate the opportunity for criminals, while enriching legitimate consumer experience. By using collective intelligence to fight back, governments, banks and businesses can not only reduce financial losses but, more importantly, restore public trust."

Martins said that criminals have extensive knowledge of social engineering, banking processes and money movement, from converting stolen funds into crypto to routing them through mule accounts. "None of those financial institutions have that kind of visibility," he said.

Building on its early research work with Nestor and the patent it acquired for its neural networks for transaction monitoring, ACI Worldwide has established a solid foundation in network intelligence, based on next-generation artificial intelligence. "It's not about learning from the history. It's about learning in real-time with the context of what's going on," Martins said.

As financial institutions refuse to share transactional data among themselves, ACI had to acquire or develop a capability called federated machine learning that can capture the learnings from the bank where the transaction was initiated and the one where it was received. "Since the banks are not sharing any data, this has empowered them to collaborate in a way they were not able to do so before."

ACI uses its patented Signals technology for network intelligence. Martins said network intelligence is embedded in financial messaging and fraud can be identified and flagged in real time. He claims the accuracy is so good that it drastically reduces the number of false positives. This is happening at scale, in an autonomous manner, with tens of thousands of transactions occurring per second.

Signals uses ISO 20022, an international standard for financial messaging. It uses XML syntax for more detailed transaction information, improving efficiency, compliance and fraud prevention. It is expected to replace older messaging formats, such as SWIFT's MT series.

Martins said Signals could be incorporated in India's unified payments interface, or UPI, with talks currently underway with the nation's financial regulator, the Reserve Bank of India, with whom it has worked before. To date, India's UPI has been adopted by seven countries. UPI logged 18.395 billion transactions, as of June 2025. He also said ACI is in talks with SWIFT.

Martins said the insights provided by Signals could be used by other business units such as marketing and sales.

Since it captures customer data in a federated manner, banks have access to rich insights about customer behavior. Those insights could be harnessed by other departments, such as marketing and sales. Internally, these insights can be used for processes, including underwriting, know your customer formalities and customer onboarding.

"We're collecting that context in a similar methodology that can be leveraged across the organization," Martins said.

These contextual insights could also be used by entities across the ecosystem. For example, banks and merchants that have close relationships could use Signals to do joint campaigns - to provide consumers with better experiences, better services and better offers, and all this without exchanging any data.

Signals might benefit a wider ecosystem of financial institutions, but it may take them longer to adopt it since it is developed by a single company; regulators and the competition commission may not favor that.

To get around this hurdle and hasten adoption, ACI decided to make its network intelligence technology open source. It is in discussions with global banking regulators and banks.

"ACI [on its own] will never be able to deliver all those use cases on our own. It needs to work with industry. We are providing the platform, and the ecosystem of players can build on top of that," Martins said.