Purpose
A 2-page portfolio-level readout to the Executive Sponsor at the end of every quarter. Demonstrates the program is working with numbers, not anecdotes. Drives budget renewal, scope expansion, and exec air cover.
This is portfolio-level, not per-agent. Per-agent quarterly reviews happen separately (within the same M17 cadence). The exec readout is the synthesis.
- When you use it: Within 10 business days of every quarter end.
- Who fills it: CoE Lead drives. Builder + Finance provide data.
- Audience: Executive Sponsor (primary). CFO + Department Heads + CISO (CC).
- Length: 2 pages. Anything longer means the headline got buried.
- Output: Slide deck (Google Slides / Keynote) or PDF, distributed before the monthly 1:1 with the sponsor.
Worked example (ACME Corp Q2 2026 — 5 quarters into the program)
ACME Corp — AI Adoption Program — Q2 2026 Exec Readout
Period: April–June 2026 Author: Morteza Moradi (AI Implementation Lead) Distributed: 2026-07-08 Audience: Jane Doe (CEO + sponsor) · Bob Lin (CFO) · Pat Lee (CISO) · Department Heads · cc John Smith (GC)
Page 1 — Headline + Portfolio
Headline
The program shipped 2 new production agents this quarter (total 5 in production). Total annualized ROI now ~$96K with zero Sev-1 incidents to date. EU AI Act readiness 60%. Q3 ask: budget for AI Gateway and CoE evolution to Hybrid.
Portfolio snapshot (as of 2026-06-30)
| Agent | Status | Tier | Launch | YTD ROI | YTD incidents |
|---|---|---|---|---|---|
finance-invoice-recon | Production | Med | 2026-06-13 | $1,155 (17 days) | 0 Sev-1, 0 Sev-2 |
sales-lead-qualification | Production | Med | 2026-06-04 | $4,800 | 0 Sev-1, 1 Sev-2 (resolved) |
support-ticket-triage | Production | Med | 2026-04-15 | $11,200 | 0 Sev-1, 0 Sev-2 |
hr-policy-qa | Production | Low | 2026-03-02 | $2,400 | 0 Sev-1, 0 Sev-2 |
legal-contract-summary | Production | Med | 2026-02-18 | $4,650 | 0 Sev-1, 0 Sev-2 |
marketing-content-draft | Pilot | Med | (pilot) | — | (in pilot) |
ops-vendor-onboard | Intake | (pending) | — | — | — |
| 5 production + 1 pilot + 1 in intake | $24,205 YTD (annualized $96K) | 0 Sev-1, 1 Sev-2 |
Quarterly metrics
| Metric | Q1 2026 | Q2 2026 | Trend |
|---|---|---|---|
| Production agents | 3 | 5 | +2 |
| Quarterly ROI | $8,400 | $15,805 | ↑ 88% |
| Sev-1 incidents | 0 | 0 | flat (target) |
| Sev-2 incidents | 0 | 1 | ↑ (single case, contained, post-mortem complete) |
| Avg agent HITL acceptance | 91% | 92% | ↑ (healthy band) |
| Vendor-AI catalog refreshes completed | 1 | 1 | (on cadence) |
| Department Champions active | 3 | 4 | +1 (Ops onboarded) |
Five monitoring signals — portfolio rollup
| Signal | Portfolio status |
|---|---|
| Output distribution shift | All 5 production agents within drift bands. No alerts this quarter. |
| HITL escalation rate | All in healthy band (8–14%). One agent (sales-lead-qualification) dropped to 6% in week 7 — investigated, false-positive in drift detector, no real issue. |
| Decision audit trails | 100% completeness across all agents. EU AI Act Article 12 + Article 19 retention configured for all. |
| Cost per execution (step level) | All agents within 1.5× their evaluation baseline. Total Q2 LLM + observability cost: $1,840 (vs $1,640 Q1). |
| Exception routing | All routing rules firing as designed. |
Page 2 — Risks, lessons, asks
Q2 incidents
1 Sev-2 incident this quarter: sales-lead-qualification HITL rate fell below 70% week 5.
- Root cause: prompt drift after a manual edit not run through eval.
- Mitigation: enforced eval-on-prompt-change in CI/CD (now blocks merge without eval pass).
- Post-mortem:
templates/10-post-mortem--sales-lead-qualification--2026-05-19.md - Action items: all 4 closed within 14 days.
Risk and compliance status
| Area | Status | Notes |
|---|---|---|
| EU AI Act readiness | 60% | Annex III mapping complete. Article 9–15 obligations mapped to internal controls. Articles 11 + 12 in production for all agents. Articles 14 + 15 testing in progress (Q3). |
| NIST AI RMF self-assessment | 70% | GOVERN + MAP complete. MEASURE in progress (per-agent eval reports + monitoring signals). MANAGE drafted. |
| ISO/IEC 42001 readiness gap analysis | 40% | Clauses 4–6 complete. Clauses 7–10 in progress. Certification not yet pursued. |
| Vendor-AI inventory | Current | Refreshed 2026-06-25. 14 vendor AI features tracked. 0 high-risk vendor AI features enabled without explicit approval. |
| Shadow AI surface | Low risk | One incident discovered (a Notion AI feature enabled by default for a department) — caught in Q2 catalog refresh, risk-classified Low, approved. |
Lessons fed back to the framework
- Eval-on-prompt-change must be mandatory CI/CD gate. Q2 Sev-2 root cause. Framework v1.1 updated.
- Drift detector false-positive rate matters. Need to tune sensitivity to avoid alarm fatigue. Q3 work item.
- First Champion network handoff in Q2 (Ops Champion onboarded). Framework's "Hybrid CoE" evolution is starting to show — Ops Champion will build their first agent in Q3 without CoE doing the build.
Q3 plan + asks
Plan:
- 2 new production agents (
marketing-content-draftfinishing pilot;ops-vendor-onboardcompletes intake → build → pilot) - First quarter where a Champion (Ops) builds an agent without CoE directly building it — proves Hybrid CoE works
- Continue NIST AI RMF self-assessment (target 90% by end of Q3)
- Sector overlay: pilot the EU AI Act overlay (we have EU customers; want compliance signal for enterprise deals)
Asks of the sponsor:
- Budget: $25K for AI Gateway adoption (Portkey or LiteLLM). Centralizes LLM routing, cost controls, policy enforcement across all agents. ROI: prevents future LLM cost overruns + simplifies vendor switching. Pays back in ~6 months on cost-control alone.
- Hiring: approval to add 1 FTE Data Engineer to support the data foundation (golden sets, drift detection). Currently borrowing time from Analytics team. As portfolio grows, this becomes a constraint.
- Decision: should we pursue ISO/IEC 42001 certification? Initial gap analysis shows ~12-month path. Costs ~$60K. Value: signal to enterprise customers. Need exec view before scoping further.
Open invitations
- Next portfolio readout: 2026-10-08 (Q3 2026)
- Sponsor 1:1: weekly cadence continues
- Open invitation to any department head to walk through the framework with me — 60 min, on request
Distributed
| Recipient | Role | Distribution date |
|---|---|---|
| Jane Doe | CEO + Executive Sponsor | 2026-07-08 |
| Bob Lin | CFO | 2026-07-08 |
| Pat Lee | CISO | 2026-07-08 |
| Mike Chen | Finance Champion / Head of Finance | 2026-07-08 |
| (other department heads) | 2026-07-08 | |
| John Smith | General Counsel (cc) | 2026-07-08 |
Blank template (copy below for your quarter)
# [Company] — AI Adoption Program — [Quarter] [Year] Exec Readout
**Period:** [Quarter date range]
**Author:** [CoE Lead]
**Distributed:** [YYYY-MM-DD]
**Audience:** [Sponsor name + role] · [CFO] · [CISO] · [Department Heads] · cc [General Counsel]
---
## Page 1 — Headline + Portfolio
### Headline
> [2–3 sentence headline. Lead with the most important number. End with the ask.]
### Portfolio snapshot (as of [end of quarter date])
| Agent | Status | Tier | Launch | YTD ROI | YTD incidents |
|---|---|---|---|---|---|
| | | | | | |
### Quarterly metrics
| Metric | [Prev Q] | [This Q] | Trend |
|---|---|---|---|
| Production agents | | | |
| Quarterly ROI | | | |
| Sev-1 incidents | | | |
| Sev-2 incidents | | | |
| Avg agent HITL acceptance | | | |
| Vendor-AI catalog refreshes completed | | | |
| Department Champions active | | | |
### Five monitoring signals — portfolio rollup
| Signal | Portfolio status |
|---|---|
| Output distribution shift | |
| HITL escalation rate | |
| Decision audit trails | |
| Cost per execution (step level) | |
| Exception routing | |
---
## Page 2 — Risks, lessons, asks
### This quarter's incidents
[List Sev-1 and Sev-2 incidents with root cause + mitigation + post-mortem link. If zero, say so.]
### Risk and compliance status
| Area | Status | Notes |
|---|---|---|
| EU AI Act readiness | [%] | |
| NIST AI RMF self-assessment | [%] | |
| ISO/IEC 42001 readiness | [%] | |
| Vendor-AI inventory | [Current / Stale] | |
| Shadow AI surface | [Low / Medium / High risk] | |
### Lessons fed back to the framework
- [Lesson 1]
- [Lesson 2]
### [Next quarter] plan + asks
**Plan:**
- [Concrete items for next quarter]
**Asks of the sponsor:**
1. **Budget:** [specific number + what it buys + payback]
2. **Hiring:** [specific role + justification]
3. **Decision:** [strategic question requiring exec input]
### Open invitations
- Next portfolio readout: [date]
- [Other standing invitations]
---
### Distributed
| Recipient | Role | Distribution date |
|---|---|---|
| | | |
Usage notes
- 2 pages is the cap. If you can't fit the program's state on 2 pages, the program is either too sprawling to summarize or you're including the wrong things. Either way, force the discipline.
- Lead with the headline. Sponsor reads the first paragraph. Everything else is reference. If the first paragraph doesn't convey "the program is delivering," rewrite it.
- Always include the ask. Even if the ask is "no asks this quarter, but here's what I'm watching." Don't waste the cycle.
- Distribute before the 1:1, not at it. Sponsor reads it on their own time, comes to the meeting with questions.
- Tie incidents to action items closed. "We had 1 Sev-2; all 4 action items closed within 14 days" is what an exec wants to hear, not just "we had 1 Sev-2."
- Compliance readiness shown as percentages. EU AI Act + NIST AI RMF + ISO 42001 are long-running mappings. Show progress, not just status.
- The shadow-AI line matters. It demonstrates the vendor-AI inventory is working as designed.
Common pitfalls
| Pitfall | What it looks like | Fix |
|---|---|---|
| Too long | 6–10 pages of detail | Cut to 2. Detail belongs in per-agent reviews. |
| No headline | Opens with "this quarter we did X, Y, Z..." | Lead paragraph must answer "is the program working? what should I think?" |
| Anecdote instead of metric | "The accountant loves it" | "5 hrs/week saved, 92% HITL acceptance, $1,155 ROI in 17 days" |
| Asks vague | "We need more resources" | Specific dollar amount + what it buys + payback timeline |
| No mention of risks | All upside, no risk surface | Include risk + compliance status — sponsor needs to see the whole picture |
| Lessons missing | Section 8 left blank | Force one lesson per quarter even if small |
| Distributed at the meeting | Sponsor sees it for the first time at the 1:1 | Distribute 24h ahead minimum |
Framework cross-references
framework.md§11.2 (quarterly review at portfolio level)framework.md§29 (ROI tracking — feeds Portfolio snapshot)framework.md§21 (5 monitoring signals — portfolio rollup)framework.md§15 (Registry — source of portfolio data)framework.md§22.1 EU AI Act Article 72 (post-market monitoring at portfolio level)framework.md§22.3 ISO/IEC 42001 Clause 9.3 (management review)workflows.mdStep 15 (Quarterly review)workflows.html→ In Action view → node M17 (Quarterly review)