On 7 October 2026, the Monetary Authority of Singapore (MAS) issued its Guidelines on Artificial Intelligence Risk Management for Financial Institutions. The Guidelines apply to all financial institutions (FIs) and all forms of AI, from machine learning models to Generative AI and AI agents, and set out MAS’s supervisory expectations for managing the risks that come with them.
The approach is proportionate: FIs can tailor their controls to the scale and nature of their AI use. But the foundations are expected of everyone, and the clock has started. The Guidelines take effect on 7 October 2027.
What the Guidelines require
- Board oversight: AI risks must be explicitly addressed in the FI’s risk appetite framework, with clear roles for the board and senior management.
- Identify and inventorise: a designated control function oversees consistent identification of AI use, including AI inside material third-party services, and an accurate AI inventory.
- Risk materiality: every AI use case is assessed for impact, complexity and reliance, with controls proportionate to the result.
- Phased timeline: oversight, inventory and materiality expectations apply from 7 October 2027; lifecycle controls and capabilities by 7 October 2028.
Why “knowing your AI” is the hard part
Most FIs already use more AI than they can easily list. Models sit in credit and fraud teams, Generative AI tools are used for drafting and research, and AI features arrive quietly inside software-as-a-service products. The Guidelines expect FIs to find all of it, including shadow AI and AI embedded in third-party services, and to keep the inventory current as use cases are added, updated or decommissioned.
Each entry then needs a defensible risk materiality rating, and a clear link to the controls applied, its owners and supporting documentation. AI use that is low risk can follow basic governance, but only if the FI can show why it meets that threshold, and it must review that decision periodically. In our research with The Hong Kong Chartered Governance Institute (HKCGI), we found that while AI adoption is accelerating, structured governance, oversight and lifecycle controls often have not kept pace. The MAS Guidelines turn that gap into a supervisory expectation.
Turning the Guidelines into an operating framework
This is where tools like Diligence can help. Diligence is Wizpresso’s AI-powered governance, risk and compliance platform, designed to turn regulatory text into structured requirements and connect them to evidence.
- Map every expectation: upload the Guidelines and Diligence extracts the requirements into an editable checklist, classifying obligations, guidance and definitions so teams can see what applies to them.
- Find the gaps: assess existing model risk, technology risk and outsourcing policies against each requirement, marked Fulfilled, Partially Fulfilled or Not Fulfilled, with rationale, citations and page references.
- Prepare for both milestones: assign owners to each gap, track policy versions, and produce reports showing the board how the FI is progressing towards October 2027 and October 2028.
Building and maintaining the AI inventory is a second, operational challenge. Adnoto, Wizpresso’s AI vetting and case-management platform, can turn existing AI registration forms into structured intake, assess each submission against the FI’s risk materiality criteria, and route ratings to the designated control function for approval, with a complete audit trail of every change and decision.
In both cases, AI does the reading and the first assessment, and the FI’s own professionals review, override and approve every result.
Governance that keeps pace with adoption
MAS has been clear that AI adoption and its risks will keep evolving, and that FIs should regularly review whether their risk management keeps up. Institutions that build their AI inventory, materiality assessments and evidence trail now will be able to adopt new AI with confidence, rather than catching up when supervisors ask.
In Singapore, Diligence and Adnoto are available through our exclusive reseller, WizReg Technologies.
👉 Learn more about Diligence and request a demo: https://wizpresso.com/products/Diligence
Read MAS’s Guidelines: https://www.mas.gov.sg/regulation/guidelines/guidelines-on-artificial-intelligence-risk-management-for-financial-institutions