Solytics Partners’ AI & MRM Conclave - Singapore 2026

Date: Monday, 23 Nov 2026‍
Time: 01:30 PM to 06:00 PM SGT‍
Cocktails & Dinner: 06:00 PM to 09:00 PM SGT
Venue: TBD

Join industry leaders, top AI executives, and AI governance and model risk professionals for an exclusive event examining how the new MAS AI Risk Management Guidelines are shaping AI governance and Model Risk Management across financial institutions in Singapore.

All times are in SGT (Singapore Local Time)

Guest registration and networking over welcome refreshments

• Welcome participants and introduce the objectives of the conclave
• Opening the day’s discussions on AI governance and MRM in Singapore

Topic: Global Perspectives and Leading Practices in AI Risk Management

• From model risk to agentic risk: What changes when AI stops recommending and starts acting on its own in payments, credit approvals, claims and client advice.
• Singapore’s three layers of AI governance: The final MAS Guidelines on AI Risk Management (issued 7 October 2026), the Project MindForge toolkit, and SAFR for runtime safeguards on AI agents.
• A global common core, and what comes next: How MAS, RBI, PRA, OSFI and the EU AI Act converge on inventory, risk-based tiering, independent validation and human oversight, and why boards and supervisors will expect evidence of control by October 2027, not just policy documents.

Topic: Implications of the New MAS AI Risk Management Guidelines

• Intent and proportionality: Why MAS built AI risk management on existing risk frameworks, and what changed between the December 2025 consultation and the final text.
• Where AI governance sits: Inside model risk management, alongside it or above it, and how the Chief AI Officer relates to the CRO.
• Board oversight and materiality: What a board-level AI risk appetite looks like, and how a GenAI copilot is assessed versus an agent that approves credit or moves money.
• Validation and third-party accountability: Validating LLMs and agents with no ground truth, and staying accountable when a vendor changes the model underneath you.
• Global groups and the road to October 2027: Reconciling MAS with BNM, RBI, PRA and the EU AI Act, and the evidence boards and supervisors will expect by the deadline.

Topic: One Ecosystem for Model, AI and Agentic Risk

• One inventory, four very different assets: Bring a PD scorecard, an ML fraud model, a GenAI RAG copilot and a credit-limit agent into a single inventory, including vendor and embedded AI, and assess each on impact, complexity and reliance.
• Validate and monitor what each asset needs: See scorecard and ML validation alongside GenAI tests for hallucination, jailbreaks and prompt injection, then track drift, bias and hallucination in production with change control for prompts and retrieval sources.
• Live agentic runtime assurance: Watch the credit-limit agent propose actions that are checked for identity, authority and limits, then auto-executed, denied, escalated or observed, with a kill switch triggered live.
• From audit log to evidence pack, then your own readiness: Generate a regulator-ready evidence bundle in one click, then complete the self-assessment and discuss it at your tables.

Topic: The Convergence of Model Risk Management and AI Risk Management

• Governance built into the pipeline: Moving from review committees to release gates, policy-as-code and evidence the platform generates by itself.
• Change control beyond model weights: Versioning prompts, retrieval sources, tools and model endpoints, and deciding which changes trigger revalidation.
• One view of production health: Tracking drift, hallucination, bias and cost together, and when LLM-as-judge counts as acceptable evidence.
• Runtime safeguards for agents: SAFR’s checkpoint pattern, where every action is checked and then auto-executed, denied, escalated or observed, backed by a tamper-evident audit log.
• Identity, authority and vendor assurance: Who holds the authoritative agent registry, how agent mandates mirror staff authority, and what evidence AI providers should give banks.

The day brought together Singapore’s final AI risk rules, the leaders turning them into one operating model, a live walkthrough of four assets across their full lifecycle, and the engineering controls that keep GenAI and agents in check at runtime. The takeaway is clear: models, AI and agents are best governed as one ecosystem, with controls built into the platform and evidence ready well before the October 2027 and 2028 milestones. Join us for an open Q&A with our speakers and leave with practical next steps for your own organisation.

Continue the conversation with fellow speakers, risk leaders, and governance professionals over dinner and drinks.

• Gain global perspectives and leading practices in AI risk management from Dr. Agus Sudjianto, Senior Advisor at Solytics Partners and former Head of Model Risk at Wells Fargo.
• Understand the implications of the new MAS AI Risk Management Guidelines for financial institutions in Singapore.
• Explore how Model Risk Management and AI Risk Management are converging, and what this means for your governance frameworks.
• Take part in a hands-on workshop led by Solytics Partners experts.
• Connect with MRM leaders, risk heads, and AI governance experts across regulated financial institutions.

Limited seats available. Confirm your participation today: Register Now!