Fraudulent transactions across digital banking and money-transfer services are expected to exceed 2.2 billion globally by 2031, rising from 773.7 million in 2025.
The Juniper Research forecast represents growth of more than 180% over the six-year period and reflects a changing fraud landscape in which attackers increasingly target bank customers and authenticated payment journeys rather than attempting to compromise banking infrastructure directly.
Juniper Research said artificial intelligence is reducing the cost of producing personalised social-engineering attacks at scale. Generative AI can help fraudsters create more convincing messages and adapt their approaches, making individual customers more commercially attractive targets.
The emergence of agentic AI could intensify the threat. According to the research, autonomous or semi-autonomous systems may allow fraudsters to coordinate multi-stage attacks and alter their tactics dynamically as victims or security controls respond. This could make scams more persistent, personalised and difficult to identify.
“Coupled with a reduced cost to commit fraud, AI enables fraudsters to create and adapt attacks faster, while instant payments reduces the window that banks have to identify suspicious behaviour,” said Shane O’Sullivan, the report’s author.
He said banks should combine behavioural, identity and payment intelligence before authorising transactions, rather than relying primarily on controls applied after a payment has started.
The recommendation reflects the growing difficulty of stopping fraudulent transfers once funds have moved through instant-payment systems. Traditional monitoring approaches that assess individual transactions may fail to identify a broader pattern involving a compromised identity, unusual behaviour and a payment that appears technically authenticated.
Juniper Research said banks and financial institutions should build a continuous view of customer risk across the payment journey. This would involve analysing how behavioural, identity and payment signals develop together, enabling earlier intervention before a transaction is authorised.
Juniper Research’s wider market estimates indicate that spending on fraud prevention by banks and financial institutions could rise from US$36.2 billion in 2026 to US$66.2 billion in 2031, representing growth of 82.9%.
The report’s central message is that fraud prevention will need to move beyond isolated transaction checks. As AI gives criminals greater speed and personalisation, financial institutions will need to connect identity, behavioural and payment intelligence while preserving sufficient controls to prevent legitimate customers from being blocked.
