Financial institutions are preparing to spend more on AI-powered fraud prevention, but many are still struggling with the data systems their fraud teams depend on. Feedzai, a fraud-prevention technology company, found that 96% of surveyed leaders expected their budgets for AI and financial-crime technology to increase over the next two years.
Nearly three-quarters of respondents, 73%, said fraud analysts spent too much time manually gathering information from different systems. Only 41% reported fully connected data across fraud, anti-money laundering, compliance and cybersecurity teams.
For an analyst investigating a suspicious payment, disconnected systems can mean checking several sources before getting a complete picture of the customer. The problem was particularly relevant in North America, where financial institutions led the surveyed regions in AI maturity but scored lowest in data infrastructure among their own capabilities.
Banks rate their readiness higher than their capabilities
The gap between confidence and implementation was another major finding in Feedzai’s “Confidence & Capability: The Feedzai Fraud Readiness Index 2026.”
Across the 12 countries surveyed, financial institutions rated their readiness to fight fraud at an average of about 80 out of 100. A separate score based on capabilities they reported having in place came to about 67, leaving an approximate 13-point difference.
The two figures measured different aspects of readiness. The higher score reflected respondents’ own assessment, while the lower one was calculated from their answers about data infrastructure, AI maturity, intelligence sharing and operational structure. Neither was an independent technical audit of the institutions.
The findings came from a survey of 1,030 senior fraud and financial-crime leaders conducted between July 13 and Aug. 4, 2026. The research covered 12 countries, including the U.S. and Canada. North America’s overall readiness score was 70 out of 100.
Banks weigh agentic AI and intelligence sharing
Financial institutions were also looking at more autonomous fraud-prevention tools, despite concerns about how much decision-making they could safely delegate to AI.
Feedzai found that 68% of institutions surveyed were either using or planning to use agentic AI across at least three fraud use cases. Trust in autonomous decision-making remained respondents’ leading concern about the technology.
The research also found considerable interest in sharing fraud information across institutions. Some 87% of respondents said they would consider securely sharing fraud intelligence with other financial institutions, including competitors.
Feedzai linked that willingness to the threat from scams, which it identified as the largest source of fraud losses in 10 of the 12 countries studied. The company said it compared findings on fraud losses with data from its platform, which processes more than 120 billion transactions annually. The readiness scores, however, were based on survey responses.