ProSight Financial Association and Oliver Wyman, the consultancy owned by Marsh, published the 2027 ProSight Risk Outlook Survey on 10 November 2026, finding that 78% of 160 banking leaders agree regulatory requirements have become less prescriptive about how risk processes are performed, and that 87% see the shift as an opportunity to optimise them. The ProSight Risk Outlook also found 64% of institutions already use AI in risk management and 65% expect regulatory pressure on AI practices to intensify. Fintechbits found no earlier coverage of ProSight research in its archive. The 2026 edition of the ProSight CRO Outlook Survey, also prepared with Oliver Wyman, drew 142 responses, and 74% of respondents put technology and cyber risk among their top risk categories.

What the ProSight Risk Outlook found

According to the ProSight release, technology and cyber risk lead banks’ concerns at 64%, followed by fraud, financial crime and data. Transformation efforts mostly target the same areas: technology and cyber risk, operational resilience and financial crime.

Banks are investing in credit capabilities before any significant deterioration, and 39% think AI could transform credit risk within two to three years through faster underwriting and continuous portfolio monitoring. The main goals of change are responding faster to risk events and improving the quality of risk management; cost reduction ranked behind both.

On AI, 64% already use it within risk and another 26% are running pilots. Many describe their AI risk appetite, governance and controls as only somewhat developed. Data is the biggest obstacle, with 49% ranking data quality and availability among their top three barriers.

Jen Guidi, chief executive of ProSight, said: “The shift in regulatory tenor has given risk leaders much more room to ask: ‘Is this the best way to do this work?'”

The survey covered leaders across both risk and business lines, which matters because the redesign banks describe would move risk decisions closer to the businesses that take the risk. Respondents’ main goals were faster responses to risk events and better quality, with lower cost a secondary aim.

The supervisory backdrop

US bank supervision has shifted toward focusing on material financial risks rather than process and documentation, and agencies have revised several guidance documents and examination practices. Banks read that as room to redesign controls built up after the 2008 crisis.

The survey shows banks plan to use that room for AI, while expecting AI itself to become a supervisory focus. Fintechbits has covered the related shift in AI for AML compliance and third-party risk under DORA in Europe.

Fintechbits analysis

Why the ProSight Risk Outlook Points to a Governance Gap

The most telling ProSight Risk Outlook finding is the gap between adoption and governance. Two-thirds of banks use AI in risk, a further quarter are piloting it, and many say their controls over it are only somewhat developed. That is the pattern that tends to end in enforcement actions when a model fails.

Less prescriptive rules cut both ways. They let banks redesign risk work, but they also shift responsibility to the bank to show its approach is sound. A supervisor that no longer specifies how to do something will still ask why a bank chose its method when something goes wrong.

For fintech vendors selling to banks, the findings are encouraging. Banks say they want to move faster and improve quality, not just cut costs, and data quality is the top barrier. Vendors that solve data problems, rather than offering AI models alone, are best placed. Oliver Wyman, a Marsh business, has a commercial interest in that advisory work.

Credit is the area to watch for change. Banks say they are investing ahead of any downturn, and 39% think AI could transform credit risk within two to three years. Continuous monitoring of borrowers, rather than annual reviews, is the most likely first use, because it uses data banks already hold.

Frequently asked questions

What is the ProSight Risk Outlook?

The 2027 ProSight Risk Outlook Survey, published on 10 November 2026 with Oliver Wyman, surveyed 160 banking leaders across risk and business lines about risk priorities, transformation and AI.

How many banks use AI in risk management?

64% of institutions surveyed already use AI within risk and another 26% are running pilots, according to the ProSight Risk Outlook.

What is the biggest barrier to AI in bank risk?

Data quality and availability, ranked among the top three obstacles by 49% of respondents.

What to Watch Next

The signal to watch is supervisory guidance on AI model risk. If US regulators issue updated expectations, the 65% of banks in the ProSight Risk Outlook who expect more AI scrutiny will be proved right, and governance spending will rise.