Artificial intelligence (AI) is rapidly transforming banking across credit, treasury, and enterprise risk.
According to a new survey by FT Longitude, SAS, and Microsoft, AI adoption has reached near-universal levels at financial institutions, supporting these organizations in their decision-making.
The research, which polled 300 credit, treasury and risk leaders at banks across 31 countries in June and July 2026, found that 82% of banks now use generative AI (genAI), and 41% use agentic AI in their decision-making processes. When combined with traditional, predictive AI such as machine learning (ML), 299 of the 300 banks use AI in credit and risk choices, underscoring a swift rise in adoption.
These levels reflect more than casual use. Some 59% of respondents have moved beyond experimentation, with 44% reaching partial, and 15% widespread embedding of the technology within decision-making. This extent of use is consistent across credit, treasury, and enterprise risk areas.
A breakdown of AI adoption across functions reveals that, with just one exception, every bank surveyed is using AI to some degree in half or more of these banking processes. On average, across the main credit risk processes, 63% had partially or widely embedded AI, while 55% did so in asset liability management and risk steering.
AI is also influencing decisions related to credit and capital management areas. For over 50% of respondents, AI is now shaping or making credit-related decisions in origination, customer management, as well as collection and recoveries. However, the technology is less frequently involved in capital and portfolio composition, yet over four in ten banks surveyed still use it in this capacity.
Governance lags
The study also found that while AI adoption continues to gain speed, governance is falling behind. Only 18% of banks currently maintain fully operational, embedded AI governance processes, and 26% lack a dedicated AI model risk policy. A mere 19% have real-time AI model drift monitoring.
Furthermore, 42% of banks lack defined “kill switches” to deactivate AI models that cross preset boundaries, and only 36% continuously retrain or update AI models.
The report also highlights key obstacles banks face when expanding automated decision-making, emphasizing internal resistance, trust, and data quality concerns.
In particular, 71% of respondents reported that internal resistance to AI led their bank to limit its use in automated decision-making. Lack of trust in data quality was cited by 79% as a primary or partial factor in restricting autonomous AI use.
Uncertainty or doubt around model risk and performance also contributes to restrictions on autonomous AI use for decision-making. 77% of respondents identified regulatory and compliance risks as a main or contributing cause for their restrictions.
Automated decision making on the horizon
Although some banks already allow AI to influence their decision making, most remain unwilling to release full control. At present, only a few institutions permit independent AI decisions, with just 10% reporting that AI makes autonomous credit or risk decisions at scale today.
According to the report, this stems from current policy. Now, 70% of banks place moderate or more extensive limits on the use of AI in autonomous decision-making.
However, future expectations diverge markedly. 76% of respondents foresee AI making autonomous credit and risk decisions at scale by 2031, and 25% expect this shift within two years.
Across credit functions, the most eagerly expected gain in value from AI in lending over the next two years is credit-decision optimization, cited by 75% of respondents. Similarly, 58% predict that AI will bring benefits through automation and optimization of decision strategies.
AI adoption in Singapore
Financial institutions in Singapore rank among the most AI‑ready and technologically advanced in the world. According to the Financial Services State of the Nation Survey 2026, conducted in November 2025, 64% of institutions in Singapore were already actively deploying AI across key business functions. A further 35% were piloting or researching AI, demonstrating sustained investment and a healthy innovation pipeline. None reported having no plans to adopt AI.
To further boost the use of new financial technologies such as AI, and catalyze innovation in the financial sector, the Monetary Authority of Singapore (MAS) launched this week the Future of Finance Institute (FFI).
The group intends to collaborate with industry to accelerate the transition from experimentation to broad-based deployment, offering capabilities including a repository of validated use cases, deployment playbooks and solution providers; implementation toolkits; and industry sandboxes to test innovations in a safe and controlled environment.
FFI is starting with AI in finance, bringing financial institutions together to share practical experience in implementing AI, develop good practices for deploying agentic AI, and strengthen the management of risks from AI use. It will progressively expand into other areas of financial innovation, including tokenization.
Featured image: Edited by Fintech News Singapore, based on image by 21studio via Magnific
