This matrix defines who is Responsible (does the work), Accountable (owns the outcome and signs off), Consulted (provides input before the decision), and Informed (notified after the decision) for 18 key program activities across 10 governance roles. It operationalizes the decision rights defined in the Governance Model §4 and the authority hierarchy in the Program Charter §11. Red-bordered cells indicate blocking authority — the activity cannot proceed without that role's approval.
| Activity | Exec Sponsor (M. Kavanagh) |
Program Dir (C. Tyrrell) |
PMO Lead (T. Valdez) |
AI Gov Dir (S. Khurana) |
Validation Lead (P. Okafor) |
Chief Architect (D. Chen) |
CISO (M. Hassan) |
VP Compliance (J. Martinez) |
Gen Counsel (R. Thorne) |
BRD Lead (per BRD) |
|---|---|---|---|---|---|---|---|---|---|---|
| Program Authorization & Governance | ||||||||||
| Charter approval & baseline | A★ | R | C | C | I | C | I | C | C | I |
| Phase-gate sign-off | A★ | R | R | A★ | C | A★ | C | C | I | C |
| Governance Model maintenance | A | R | C | R | I | C | I | I | C | I |
| Scope & Requirements | ||||||||||
| JAD session facilitation | I | R | R | R | I | C | C | C | C | R |
| Requirements sign-off | I | I | R | A★ | I | A★ | A | A | A | A |
| Scope change approval | A★ | R | C | C | I | C | I | I | I | C |
| AI Model Governance | ||||||||||
| Model design concept approval | I | I | I | A★ | C | C | C | I | I | R |
| Independent model validation | I | I | I | C | A★ | I | C | I | I | I |
| Model production release | I | C | I | A★ | A★ | C | C | I | I | R |
| Fairness / bias testing | I | I | I | A | R | I | I | C | C | C |
| Shadow AI disposition | C | C | I | A★ | I | C | I | I | I | I |
| Technical & Security | ||||||||||
| Architecture design approval | I | C | I | C | I | A★ | C | I | I | R |
| Security architecture sign-off | I | I | I | C | I | C | A★ | I | I | C |
| Emergency production halt (AI incident) | I | I | I | R | C | I | R | I | I | I |
| Financial & Resource | ||||||||||
| Budget change >$500K | A★ | R | C | I | I | I | I | I | I | C |
| Contingency drawdown >$1M | A★ | R | C | I | I | I | I | I | I | I |
| Vendor contract >$500K | A★ | R | C | I | I | C | I | I | C | I |
| Regulatory & Compliance | ||||||||||
| Regulatory compliance interpretation | I | I | I | C | I | I | I | A | R | I |
Reading This Matrix
Blocking Authority (A★)
Red-bordered cells marked A★ indicate blocking authority — the activity literally cannot proceed without that role's approval. This is the matrix's most important feature. Three examples:
- Phase-gate sign-off requires three concurrent A★: Executive Sponsor, AI Governance Director, and Chief Architect must all approve. One board cannot override another.
- Model production release requires two A★: Both the AI Governance Director AND the Independent Validation Lead must approve. Neither can be overridden — not by the Executive Sponsor, not by schedule pressure, not by any other authority. This is the two-line-of-defense structure made explicit.
- Shadow AI disposition requires one A★: Only the AI Governance Director can approve the disposition (fold-in, defer, or reject) of an identified shadow AI initiative.
The "R but not A" Distinction
The Program Director (C. Tyrrell) is R (Responsible — does the work) on many activities but A (Accountable — owns the outcome) on very few. This reflects the matrix management reality: the Program Director coordinates, prepares, and presents — but the governance boards own the decisions. The Program Director prepares the phase-gate package; the three boards decide whether the gate passes. The Program Director prepares the change request; the Executive Steering Board decides whether to approve it.
Emergency Production Halt
Row 14 (Emergency production halt) shows both the AI Governance Director and the CISO as R — either can independently trigger a production halt for an AI model incident. This is deliberate: an AI hallucination incident (S. Khurana's domain) and a security breach (M. Hassan's domain) both require immediate response, and neither should have to wait for the other's availability.
Cross-References
- Program Governance Model §4 — Decision Authority Matrix (narrative version of this RACI)
- Program Charter §8–11 — Governance structure and Program Director authority
- Program Management Plan §18 — Decision rights in operational context
- AI Governance & JAD Charter §7 — JAD mandatory roles (maps to Requirements sign-off row)