State of AI in Finance & Banking 2026

Generative AI crossed from experiment to default infrastructure in banking during 2025: a Temenos survey of more than 400 banks worldwide found 54% had genAI either implemented or in active deployment, and 80% believe institutions without AI will fall behind. The story now is less about whether to adopt and more about governance, spend, and proving revenue.

Adoption is near-universal at the top, uneven at the bottom

Bank size remains the sharpest dividing line in adoption. Among institutions with over $250 billion in assets, 79% report generative AI live or in the pipeline, and 75% of banks in the $50-250 billion bracket say the same. Smaller institutions lag badly: only around 40% of banks under $10 billion in assets have reached the same stage. This gap matters because the cost of building governance, talent, and data pipelines does not scale down neatly, leaving community and regional banks at risk of a widening competitive disadvantage.

Spend is compounding faster than almost any prior banking tech wave

The financial commitment behind these pilots is enormous. Research and Markets projects total bank spending on generative AI will grow from $5.6 billion in 2024 to $85.7 billion by 2030, an increase of roughly 1,430% over the period. That trajectory implies a compound annual growth rate above 55%, far steeper than historical core-banking or cloud-migration cycles. The scale signals that boards now treat genAI as a strategic line item rather than an innovation-lab experiment, with budgets shifting from proof-of-concept to enterprise rollout.

Customer-facing use cases lead, but productivity is catching up

When banks explain why they are deploying generative AI, the answers cluster around the customer. In the Temenos survey, 64% cited improving customer experience as the primary driver, 58% pointed to enhancing customer service functions, and 55% aimed to lift internal productivity. The closeness of these figures is telling: institutions are no longer choosing between front-office and back-office value but pursuing both simultaneously. Agentic workflows are the next frontier, with 60% of respondents expecting human employees to work alongside autonomous AI tools.

Governance is forming, but uneven and largely informal

Oversight structures are emerging unevenly. Only 42% of surveyed banks reported a dedicated internal group overseeing generative AI implementation, and just 8% had created a chief AI officer role, with none of those in the United States. At the same time, virtually all banks said it took less than 12 months to obtain internal approval for genAI projects, suggesting speed is currently outpacing formal accountability. As regulators sharpen expectations around model risk and explainability, this governance gap is likely to become the defining operational challenge of 2026.

Najczęstsze pytania

More reports

Compiled by ToolGlance from publicly reported data; figures link to their sources. Updated 2026-05-30.