An interactive study of IMDA's Agentic AI governance framework
Interactive study notes on the IMDA Model AI Governance Framework for Agentic AI
(v1.5) — dense governance prose rendered as readable, interactive topics, with every
claim traceable to the source.
The current focus is roles & responsibilities: turning the framework into
a practical, calibratable RACI. Other facets of the framework may follow.
Roles & Responsibilities
The current study track — three topics building to a RACI a company can adapt.
Definition & Components
What agentic AI is, the eight core components, multi-agent patterns, and the action-space vs autonomy axes — with an interactive component diagram, a pattern switcher, and a 2×2.
Read → TOPIC 02Parties & Value Chain
The agentic value chain, internal team responsibilities, external-party handling, end-user archetypes, and where accountability sits — with an interactive value-chain diagram and team selector.
Read → TOPIC 03RACI Responsibility Assignment
A calibratable RACI across the framework’s four governance dimensions — an interactive, filterable matrix (the paper has none; the whole thing is a template, not policy).
Read → TOPIC 04RACI — FI calibration
A worked financial-institution calibration of the scaffold — IMDA’s four teams split across three lines of defence, with the single A distributed to whoever has the most control.
Read →All four R&R topics are complete — Topic 03 is the IMDA baseline; Topic 04 a worked FI calibration.