6.1 Portfolio Governance: The AI Council
As organizations scale beyond a single AI Innovation, they need a coordinating body that provides oversight without undermining pod autonomy. The AI Council serves this function—setting standards, resolving cross-pod issues, and maintaining the enterprise-wide view of AI risk and opportunity. Done right, the Council accelerates pods rather than constraining them.
The AI Council is not a steering committee that approves every decision. It's a governance body that sets boundaries and provides air cover. Pods operate autonomously within those boundaries. The Council intervenes only when pods exceed their authority, when cross-pod coordination is needed, or when enterprise-wide issues arise.
The AI Council's Role
What the Council Does
Standard Setting
Establishes enterprise-wide AI governance standards, guardrails, and requirements that all pods must follow.
High-Risk Oversight
Reviews and approves Tier 3 AI products. Provides additional scrutiny for consequential applications.
Dispute Resolution
Resolves disagreements between pods and Ethics Liaisons, or between pods competing for resources.
Portfolio Management
Maintains visibility into all AI products, identifies overlaps and gaps, and prioritizes investments.
What the Council Does NOT Do
- Approve every deployment: That's the pod's decision within their tier
- Dictate technical approaches: Pods choose their own methods
- Manage pod resources: STOs manage their own teams and budgets
- Review every Model Card: Only for Tier 3 and escalations
- Make product decisions: STOs own their product direction
Council Composition
Core Members
| Role | Responsibility | Commitment |
|---|---|---|
| Chair (CAIO or delegate) | Leads meetings, drives decisions, represents to executives | 20-30% of time |
| Chief Ethics Officer | Ensures ethical standards, manages Ethics Liaison pool | 20% of time |
| Legal Representative | Advises on regulatory compliance and legal risk | 10-15% of time |
| Security Representative | Ensures AI security standards and data protection | 10% of time |
| Technical Lead | Sets technical standards, advises on feasibility | 15-20% of time |
| Business Representative | Represents business priorities and value perspective | 10% of time |
Extended Members
Participate as needed based on agenda:
- Pod STOs: Attend when their products are discussed
- Domain Experts: For specific regulatory or technical topics
- External Advisors: For independent perspective on complex issues
- HR Representative: For workforce and AI impact discussions
Council Operations
Meeting Cadence
| Meeting Type | Frequency | Purpose |
|---|---|---|
| Regular Session | Bi-weekly (2 hours) | Charter approvals, escalations, policy decisions |
| Portfolio Review | Monthly (3 hours) | Overall AI portfolio health, trends, investments |
| Strategy Session | Quarterly (half-day) | Long-term direction, major policy changes, external trends |
| Emergency Session | As needed | Sev-1 incidents, urgent compliance issues |
Decision Making
The Council aims for consensus but has clear escalation paths:
Consensus Preferred
Most decisions should achieve consensus through discussion. Dissent is documented but not blocking.
Chair Decides
If consensus not reached, Chair makes the call after hearing all perspectives.
Executive Escalation
For decisions with major business impact, escalate to C-suite or Board as appropriate.
Standard Agenda Items
- Incident Review: Any Sev-1/Sev-2 incidents since last meeting
- Charter Approvals: New Tier 3 products, tier changes
- Escalations: STO/Liaison disputes, guardrail overrides
- Policy Updates: Changes to standards or guardrails
- External Updates: Regulatory changes, industry developments
- Open Forum: Issues raised by any member
Portfolio View
AI Product Registry
The Council maintains a comprehensive registry of all AI products:
| Registry Field | Purpose |
|---|---|
| Product Name & ID | Unique identifier for tracking |
| Owning Pod & STO | Clear accountability |
| Risk Tier | Governance requirements |
| Lifecycle Stage | Current phase (dev/prod/retiring) |
| Key Metrics | Performance and governance health |
| Model Card Link | Full documentation access |
| Last Review Date | Governance currency |
Portfolio Health Dashboard
The Council monitors aggregate health metrics:
Portfolio Prioritization
When resources are constrained, the Council helps prioritize:
- Strategic Alignment: How well does this support enterprise strategy?
- Business Value: What's the expected ROI?
- Risk Profile: What's the governance burden?
- Capability Building: Does this develop strategic capabilities?
- Dependencies: What does this enable or depend on?
The Enabling Council
The best AI Councils are measured not by what they block, but by what they enable. A well-functioning Council removes obstacles for pods, provides air cover for responsible risk-taking, and creates consistency that builds trust. When pods see the Council as helpful rather than hindering, governance becomes a competitive advantage rather than overhead.