80% Financial Institutions Faced Attacks Due To The Rise Of Agentic AI: Survey

Around 80 per cent of financial institutions globally have already encountered attacks involving agentic artificial intelligence (AI), according to a new survey by fraud prevention firm BioCatch. The survey titled, The Future of Digital Trust: AI agents and the speed of fraud and financial crime, has pointed to growing concerns within the banking sector about the role of AI in increasing the speed, scale and sophistication of financial fraud.

Fraud Attempts On The Rise

The study has found that 81 per cent reported a year-on-year increase in fraud attempts at their organisations, up from 71 per cent in 2025. At the same time, 76 per cent said fraud losses had also increased over the past year, compared to 59 per cent in the previous survey.

Respondents identified scam attempts through instant payment platforms as the biggest contributor to rising fraud activity, followed by the adoption of AI and automation tools and a rise in authorised fraud and scam attempts.

More than three-quarters, or 76 per cent, have said that they were highly concerned about the increasing speed of fraudulent activities in their regions. Nearly six in 10 respondents also fear that criminal organisations are evolving faster than financial institutions.

AI Emerging As A Key Fraud Challenge

According to the study, 88 per cent of banking professionals believe AI has increased the sophistication of fraud and scam schemes. Deepfake-enabled impersonation scams, automated phishing campaigns and automated money laundering activities were among the most commonly reported AI-enabled threats during the past year.

Around 50 per cent had witnessed deepfake-enabled social engineering or impersonation attacks, while 48 per cent had reported automated phishing attacks linked to AI tools.

Banking professionals also expressed concerns about the next phase of AI adoption. About 84 per cent stated that AI agents could become the industry's largest exploitable vulnerability for fraudsters within the next year.

Further, 72 per cent mentioned that distinguishing legitimate AI-assisted banking activity from malicious or manipulated activity would be very challenging in a future where AI agents regularly initiate or assist transactions.

Growing Push For Industry Collaboration

The survey highlighted that greater support and cooperation among financial institutions are the only ways to tackle fraud. About 85 per cent fear that real-time information sharing between banks could positively affect their ability to prevent fraud, scams and financial crime.

Similarly, 86 per cent said that access to real-time intelligence about recipient accounts in interbank transactions would improve their ability to identify and stop scams.

However, respondents cited barriers such as privacy regulations, legal liability concerns and data-sharing restrictions as obstacles to broader collaboration.

Nearly half of those surveyed said that they encounter fraud schemes involving international actors or cross-border elements at least once a day, while 76 per cent said such cases occur at least weekly.

Customer Trust Under Pressure

The survey also pointed to the wider impact of fraud on customer relationships. Around 74 per cent of respondents have stated that fraud and scam incidents frequently damage customer trust or engagement.

At the same time, 68 per cent of banking leaders said that their institution's approach to fraud prevention and reimbursement had resulted in a net loss of customers.

While financial institutions continue to invest heavily in fraud prevention, many are finding it difficult to keep pace with increasingly sophisticated criminal networks that are making greater use of AI-powered tools and automation.