Firms race to counter fast-evolving AI-driven fraud

A global survey of credit risk and fraud executives found that half struggle to detect and respond quickly to new fraud trends, as concerns grow about criminals using artificial intelligence to scale attacks.

The results highlight a gap between awareness and operational readiness. Respondents said fraud patterns change quickly and spread across digital channels, while many organisations still rely on static rules and fragmented data.

Speed Gap

Half of executives said their biggest challenge is detecting and reacting quickly to new fraud trends. That leaves firms exposed when schemes shift in real time and outpace established controls.

Executives also reported pressure to strengthen defences while keeping digital journeys smooth for customers. Online onboarding and application processes remain a focus because they combine high volumes with the need for fast decisions.

AI Threats

Concerns about AI-enabled fraud featured strongly. More than three-quarters of executives (77%) said they are concerned about threats that use AI.

Respondents linked that anxiety to automation and attackers' ability to mimic legitimate behaviour. They also pointed to the scale of digital processes, which can amplify losses when fraudulent activity is not detected early.

These concerns cut across sectors. Executives in financial services and insurance reported similar themes to those in retail and telecommunications, reflecting widespread digital identity and payment fraud risks.

Defence Investments

Many organisations said they are deploying AI-based tools and real-time monitoring. Three-quarters of respondents (75%) reported using AI-driven, adaptive fraud prevention solutions.

A similar share is increasing real-time monitoring for suspicious patterns, with 74% using real-time anomaly detection.

Together, these moves suggest a shift away from rule-based approaches. AI-based methods can incorporate model outputs and live behavioural signals, updating decisions as new information emerges.

Data Integration

Executives also identified the most important building blocks for a broader fraud strategy. The top priority was improved visibility across customer data, with 33% selecting a comprehensive fraud risk view of customer data.

Customer experience ranked close behind: 23% cited reducing friction as critical. Another 22% emphasised aligning data at the customer level rather than by channel.

Organisational structure and data silos also featured, with 19% pointing to breaking down silos between fraud and credit teams.

These priorities suggest fraud and credit functions remain separate in many firms, with different datasets and decision processes. A consolidated customer view remains a recurring objective, particularly for organisations that need to spot patterns spanning products and channels.

Next Priorities

Looking ahead, executives highlighted a mix of operational and technical priorities. More than half (54%) aim to improve operational efficiency and automation.

Model performance was also a leading focus. Another 54% plan to improve the accuracy of AI and machine learning models, while 52% are focused on reducing overall fraud losses.

These goals reflect a persistent trade-off between strict controls and a low-friction customer journey. False positives raise costs and disrupt legitimate activity, while missed fraud increases losses and can trigger chargebacks and customer complaints.

Business Role

The results also suggest fraud prevention is increasingly viewed as part of business performance, not just loss control. Respondents said fraud strategies are expected to support customer retention, experience, and profitability.

In the responses, 57% linked fraud strategies to higher customer retention and loyalty, while 55% said they should improve customer experience and reduce friction. Another 51% pointed to improving profitability or risk-adjusted returns.

Half of respondents also said fraud strategies are expected to enable event-driven, real-time decisioning. That points to a broader shift toward using risk signals across customer interactions, from onboarding to account management and transactions.

"The biggest risk today isn't just fraud itself, rather it's the speed and sophistication at which fraud vectors are evolving, with half of institutions indicating keeping pace with this race remains their greatest challenge," said Carol Hamilton, Chief Commercial Officer, Provenir. "AI-driven, real-time, and adaptive fraud prevention is essential to improving model accuracy and protecting customers without adding friction to the experience."

The Harris Poll conducted the survey in early December 2025, gathering responses from 203 directors and above across North America, EMEA, Latin America, and Asia Pacific. Participants had responsibility for deploying AI-based solutions, with roles spanning risk evaluation, credit approvals, fraud detection, and personalised offers across industries including banking, financial services, fintech, insurance, retail, wholesale, and telecommunications.