Banking insights: 2026 AI Impact Survey Report

Stronger AI controls enable more confident decisions

Banks are scaling AI across multiple front-, middle- and back-office functions, with top benefits including greater efficiencies (62%), improved decision-making (56%) and higher-quality outputs (42%).

But just 32% report revenue growth, and only 36% report cost reduction. More banks should be experiencing the cost-saving and profit-driving benefits of automation, but weak governance limits performance potential.

On paper, banks have governance policies in place. But at the same time, banking leaders identified governance as their top barrier to AI performance, and only 18% say they’re fully confident in their AI controls. While many organizations may wait for regulators to provide guidance before stress-testing their AI controls, that delay is preventing them from building confidence in where and how they deploy AI today.

A lack of decision rights and guardrails are common roadblocks preventing financial institutions from testing their controls and incident response. Those with strong governance and controls have clear responsibilities across the C-Suite, rather than ad hoc committees. 

Banks can also establish clearer internal guidance by forming governance models aligned with existing regulatory requirements such as SR 11-8, Basel III, the EU AI Act and upcoming U.S. state regulations.

The U.S. Treasury released guidance in February that standardizes best practices for managing AI cybersecurity risks in the financial services sector, rooted in the Financial Services AI Risk Management Framework. Financial institutions that align their AI governance policies with this framework and the six interrelated resources released by the Treasury can design and conduct assessments, address gaps, prioritize mitigation efforts and develop a more resilient control posture. Aligning their AI governance framework to this guidance provides banks with a stronger foundation for more confident AI decision-making across the enterprise. 

Banks deploy AI across a workforce that can’t use it

Many banks are still piloting AI in low-risk areas such as reconciliation processes, anti-money laundering (AML) programs and customer service enablement. But few say AI is fully embedded into operations. Nearly half of banking respondents cited insufficient training as a reason past AI initiatives have failed, and 36% identify workforce readiness as a top concern about implementing agentic AI.

That points to a broader issue with how AI is being deployed in banks. When institutions layer AI on top of existing customer onboarding, transaction monitoring or fraud detection workflows, they’ll see marginal returns. Institutions where AI is actually changing performance outcomes have redesigned processes with automation and autonomous AI in mind. That requires redesigning the overarching operating model and underlying workflows, not just deploying tools.

That process differs between large institutions and regional banks. Where large banks may struggle with changing processes across hundreds of interdependent systems, regional institutions may lack the AI and data readiness expertise required to redesign workflows and operationalize controls effectively. But with an AI strategy grounded in governance, banks of all sizes can determine their most impactful use cases and the workflows that will need to support them.

Ready to talk? We’re ready to listen.

Request a meeting -->

Banks are underusing AI to improve data readiness

Similar to other financial services industries surveyed, data readiness is a barrier to deploying AI in banking. Data barriers often trace back to inconsistent definitions across business lines, siloed systems that were not built to communicate and too many stakeholders with fragmented ownership.

Larger institutions are addressing this by building semantic layers on top of existing systems rather than centralizing data — creating a usable, logical view that makes data accessible to AI without requiring infrastructure rebuilds. Regional banks can use AI to aid in data prioritization and remediation, supporting continuous improvement in data quality and integration.

Three actions to move from AI potential to performance

In today’s financial services market, the gap between AI ambition and performance is real. And it will not be fixed until banking leaders ground AI initiatives in governance.

Grant Thornton works with companies across the financial services industry to strengthen their AI strategy and build readiness through solutions including creating a responsible AI governance framework for banks that aligns with organizational goals; building foundational data readiness; and embedding AI into workflows such as key customer onboarding, compliance processing, and fraud detection. Discover how our AI solutions help banks turn AI strategy into measurable value.

Methodology

Between Feb. 23 and March 18, 2026, Grant Thornton surveyed 950 business leaders, a group restricted to CFOs, CIOs/CITOs, COOs, and VPs, department heads, and directors who report directly to the C-suite. The banking-specific subgroup comprises 50 respondents. Role-specific findings within the banking subset of data are directional only.