AI in Banking: GenAI Survey Insights (EY)

Summary

• A total of 71% of surveyed bank respondents have implemented or soft-launched generative AI capabilities.p.6
• The proportion of banks that have fully implemented GenAI applications rose to 47% in 2025 from 10% in 2023.p.9
• 58% of respondents experienced between 6% and 20% in revenue uplift from GenAI applications.p.10
• Banks invested ~$13m in GenAI over the past two years and plan to increase investments to around $21m in the next two years.p.13
• A formal governance committee is present in 75% of banks to manage the ethical use of AI technologies.p.17
• Among implemented GenAI use cases, 40% failed to achieve planned results or were discontinued.p.12

Key insights

1. Although 99% of bank respondents are familiar with agentic AI, only 31% have pursued implementation.p.6
2. Front-office applications represent 43% of implemented GenAI use cases, whereas back-office accounts for 23%.p.6
3. Banks expect to put an average of ~15 GenAI models or use cases into production over the next two years.p.14
4. Card fraud detection is the top use case explored by retail banks at 80%, while loan servicing leads for commercial banks at 70%.p.16
5. Primary decision-making authority for generative AI initiatives rests with executive leadership teams in 60% of banks.p.17
6. More than 70% of banks report workforce reductions in customer service and operations departments following GenAI implementation.p.18
7. Commercial banks prioritize augmented AML/KYC capabilities as their top back-office GenAI investment at 72%.p.20