Executive Summary
- Most AI initiatives die not in execution but in the gap between “we should do something about AI” and the first structured conversation. This agenda closes that gap with a fill-in-the-blank template designed to be used within five business days of attending the briefing.
- The meeting has one job: produce three decisions in 60 minutes — who owns this, what gets audited first, and what does “success” mean before any money is spent. Organizations that define success metrics before launching achieve 54% pilot success versus 12% without (Pertama Partners, n=2,400+ AI initiatives, 2026).
- This agenda is built for 4-8 people in a room, not a town hall. The right attendees are the people who will make or break the first 90 days: the executive sponsor, the CFO or finance lead, the CIO or IT lead, the general counsel, and 1-2 department heads whose teams will pilot first.
Before the Meeting: Three Things to Prepare
The meeting sponsor (the person who attended the briefing) completes these before sending the calendar invite. Total prep time: 90 minutes.
1. Fill In the Current State Summary
This becomes the meeting’s opening slide or printed handout. Fill in what you know; mark what you don’t with “UNKNOWN — this is what the audit will answer.”
| Question | Your Answer |
|---|---|
| How many employees use AI tools today (sanctioned or personal accounts)? | _______ (industry average: 78% use unapproved tools — WalkMe, n=1,000, 2025) |
| What AI features are already active in tools you pay for? (M365 Copilot, Salesforce Einstein, Google Gemini in Workspace) | _______ |
| Has anyone run an AI pilot before? What happened? | _______ |
| What is the organization’s current AI spend? (Include personal subscriptions employees expense) | $_______ /month (if unknown, write “UNKNOWN”) |
| Does an AI acceptable use policy exist? | Yes / No / Partial |
| Has leadership made any public statement to employees about AI and jobs? | Yes / No |
2. Decide Who Belongs in the Room
The meeting needs decision-makers, not observers. Four to eight people.
| Role | Why They Must Be There | Name |
|---|---|---|
| Executive sponsor (you, or the CIO/COO who owns AI outcomes) | Makes final decisions. Without a named sponsor, success rate drops from 68% to 11% (Pertama Partners, 2026). | _______ |
| CFO or finance lead | Approves the real budget — license fees are 10-20% of total cost (BCG 10-20-70, 2025). No CFO buy-in, no surviving the first quarterly review. | _______ |
| CIO or IT lead | Owns the shadow AI audit, tool evaluation, and security review. Knows what is already deployed. | _______ |
| General counsel | Flags the legal exposures that exist right now — data in personal AI accounts, EU AI Act (August 2026), vendor terms. | _______ |
| 1-2 department heads (highest-volume workflows) | Business owners of the first pilot. AI deployed by IT without business ownership fails 61% of the time (Fair Observer, 2025). | _______ |
Do not invite: The entire leadership team, anyone who will attend but not act, or vendors.
3. Send the Pre-Read
Attach the “brief your boss” one-pager from the briefing. One sentence in the calendar invite: “Attached is a 3-minute summary of the AI briefing I attended. The meeting will focus on three decisions: who owns this, what we audit first, and how we define success.”
The 60-Minute Agenda
0:00–0:10 — What We Know and What We Don’t (Sponsor Presents)
Purpose: Establish a shared factual baseline. No opinions about AI’s potential. No vendor names. Just the current state.
Present the Current State Summary you filled in above. Spend no more than 5 minutes on it. Then share three numbers from the briefing:
- Only 5% of organizations generate substantial financial returns from AI (BCG, n=10,600, 2025). The failure is not technology — it is deploying without defined success criteria, honest cost modeling, or workflow redesign.
- 78% of employees already use unapproved AI tools (WalkMe, n=1,000, 2025). The question is not whether your organization uses AI. It is whether anyone governs it.
- 5% of companies capture measurable value. They do three things: deploy on proven tasks first, budget for the full cost (not just the license), and redesign workflows around where AI moves the bottleneck.
Close with one sentence: “This meeting is about putting us on the path to the 5%. We need three decisions in the next 50 minutes.”
0:10–0:25 — Decision 1: Who Owns This? (Group Discussion)
Purpose: Name a single person accountable for AI outcomes in the first 90 days. Not a committee. Not “IT.” One name.
The evidence: Projects with sustained C-level sponsorship achieve 68% success versus 11% without (Pertama Partners, n=2,400+, 2026). 56% of failed AI projects lost executive sponsorship within six months (Fair Observer, 2025). The sponsor is the single strongest predictor of whether this works.
Fill in:
| Decision | Answer |
|---|---|
| Who is the executive sponsor? | _______ |
| What percentage of their time is allocated for the first 90 days? | _______% (minimum: 15-20%) |
| Who do they report progress to, and how often? | _______ (recommended: biweekly to CEO or leadership team) |
Three models that work at 200-2,000 employees:
| Model | Best When | Sponsor |
|---|---|---|
| CIO-led | Primary use cases are technical (coding tools, IT automation, data analytics) | CIO with COO as business counterpart |
| COO-led | Primary use cases are operational (customer service, invoice processing, document routing) | COO with CIO as technical counterpart |
| Cross-functional pair | AI touches multiple functions simultaneously | CIO + COO co-own, with clear decision rights for each domain |
What NOT to decide here: Do not create a committee, a steering group, or a “center of excellence” in this meeting. Those are month-2 structures. Today, name one person. The governance architecture comes after the audit reveals what needs governing.
0:25–0:40 — Decision 2: What Gets Audited First? (CIO/IT Lead Presents Options)
Purpose: Commission the shadow AI audit and the workflow inventory that will drive every decision for the next 90 days.
The shadow AI audit answers three questions:
- What AI tools are employees already using? (Expect 3-5x the known footprint)
- What company data is flowing into those tools? (77% of employees paste data through personal accounts — LayerX, 2025)
- What is the real monthly AI spend? (Include personal subscriptions, embedded vendor features, and expensed accounts)
The audit takes 10-15 business days and costs $0 in external spend (internal labor only). It has five steps:
| Step | Days | Owner | Output |
|---|---|---|---|
| Mine expense reports for $10-50/month AI subscriptions | 1-2 | Finance | List of known AI spend |
| Pull OAuth/SSO consent logs for AI applications | 1-2 | IT/Security | List of AI tools with company credentials |
| Scan network/DNS logs for AI endpoint traffic | 2-3 | IT/Security | Traffic analysis |
| Send 5-question anonymous employee survey (“What AI tools do you use for work?”) | 3-5 | HR + IT | Self-reported usage |
| Compile findings into one-page summary: tools, data exposure, estimated spend | 2-3 | Sponsor | The baseline document |
In parallel: the workflow inventory. Ask the 1-2 department heads in the room to identify their three highest-volume, most repetitive workflows. Score each on:
| Criteria | What to Assess |
|---|---|
| Volume | 50+ instances per week? |
| Measurability | Can you measure time-per-unit or error rate today? |
| Data availability | Do digital inputs already exist? |
| Error tolerance | Is the cost of a mistake low enough for a learning curve? |
| Complexity | 5-15 steps, 1-2 handoffs? (Not client-facing, not multi-department) |
Fill in:
| Decision | Answer |
|---|---|
| Shadow AI audit start date | _______ (within 5 business days of this meeting) |
| Audit lead | _______ (CIO or senior IT leader with cross-functional access) |
| Audit completion target | _______ (10-15 business days from start) |
| Department heads completing workflow inventory | _______ and _______ |
| Workflow inventory due date | _______ (same date as audit completion) |
0:40–0:55 — Decision 3: What Does Success Look Like? (Group Discussion)
Purpose: Define success criteria before spending a dollar. This is the decision that separates the 54% success rate from the 12% (Pertama Partners, n=2,400+, 2026).
Do not define success as “adopt AI” or “improve productivity.” Those are aspirations, not metrics. Define success for the first 90 days in terms the CFO can audit:
Fill in:
| Metric | Target | How Measured | Who Owns It |
|---|---|---|---|
| Shadow AI audit completed with findings presented to this group | Done / Not Done by _______ | Written report delivered | _______ |
| Acceptable use policy drafted and reviewed by counsel | Done / Not Done by _______ | Document exists and is distributed | _______ |
| One pilot workflow selected with baseline metrics captured | Selected by _______ | Named workflow with before-measurement | _______ |
| Pilot launched with 15-25 users | Launched by _______ | Users active in tool | _______ |
| 90-day business case presented to leadership | Presented by _______ | Document with cost actuals vs. plan, measured outcomes, recommendation | _______ |
The honest cost conversation (CFO, this is yours):
The license fee — $19-50/seat/month — is 10-20% of the actual cost. The remaining 80-90% is debugging AI output, additional review time, training, governance, and workflow redesign. Year 1 total cost runs roughly 2.5x the license (DX Research/Atlan, 2025). The CFO’s job today is not to approve a budget. It is to agree that the business case will include the full cost, not just the subscription.
Fill in:
| Decision | Answer |
|---|---|
| Budget methodology | Full-cost model (2.5x license Year 1) / License-only / Other: _______ |
| Who builds the cost estimate after the audit? | _______ |
| When does the cost estimate come back to this group? | _______ |
0:55–1:00 — Next Steps and Calendar (Sponsor Closes)
Purpose: Lock the next three meetings before anyone leaves the room.
| Meeting | Date | Purpose |
|---|---|---|
| Audit readout (2-3 weeks from today) | _______ | Present shadow AI audit and workflow inventory findings. Make the pilot selection decision. |
| Pilot launch check-in (5-6 weeks from today) | _______ | Confirm pilot is live, review first-week adoption data, address friction. |
| 90-day business case review (12 weeks from today) | _______ | Present results: what was spent, what was measured, what the recommendation is for scale or kill. |
The sponsor sends a follow-up email within 24 hours containing:
- The three decisions made today (sponsor name, audit timeline, success criteria)
- The Current State Summary — updated with any new information surfaced in the meeting
- The acceptable use policy template (two pages, four sections: approved tools, data rules, human review requirement, incident contact)
- Calendar invitations for the three follow-up meetings
One thing NOT to do after this meeting: Do not start evaluating AI vendors. The audit determines what you need. Vendors are a week 3-4 conversation. The most common failure mode at this stage is buying a tool before understanding the problem.
Key Data Points
| Metric | Finding | Source |
|---|---|---|
| Organizations capturing substantial AI value | Only 5% | BCG, n=10,600, 2025 |
| Success with C-level sponsor vs. without | 68% vs. 11% | Pertama Partners, n=2,400+, 2026 |
| Success with pre-defined metrics vs. without | 54% vs. 12% | Pertama Partners, n=2,400+, 2026 |
| Employees using unapproved AI tools | 78% | WalkMe/Propeller Insights, n=1,000, 2025 |
| Employees pasting company data into personal AI accounts | 77% | LayerX, October 2025 |
| True cost vs. license fee | 2.5x Year 1 (license = 10-20% of total) | DX Research/Atlan, 2025 |
| Failed projects that lost C-suite sponsorship within 6 months | 56% | Fair Observer, 2025 |
| Failed projects treated as IT rather than business transformation | 61% | Fair Observer, 2025 |
| Organizations with no AI governance policy | 63% | IBM, n=604, 2025 |
What This Means for Your Organization
This agenda exists because the most dangerous moment in an AI initiative is the week after someone attends a briefing and comes back energized — but without a structure for channeling that energy into decisions. The data is consistent: organizations that hold a structured internal conversation within five business days of an external trigger and leave with named owners, defined timelines, and success criteria before spending are three to five times more likely to produce measurable results than those that “schedule a follow-up” and let the momentum dissipate.
The agenda is deliberately constrained to 60 minutes and three decisions. It does not attempt to solve the AI strategy question. It solves the “what do we do this month” question — which is the only question that matters until the audit data comes back. Every subsequent decision (tool selection, pilot design, budget approval, governance structure) becomes sharper once you know what AI already exists in your organization, where the highest-value workflows sit, and what “success” means in terms your CFO can audit.
If you used this agenda and the conversation surfaced questions the template did not anticipate — or if you want to pressure-test your audit plan or pilot selection before committing resources — that is exactly the conversation worth having at brandon@brandonsneider.com.
Sources
- Pertama Partners — n=2,400+ AI initiatives, 2025-2026. C-level sponsorship success rates (68% vs. 11%), pre-defined metrics impact (54% vs. 12%). Independent. High credibility. https://www.pertamapartners.com/insights/ai-project-failure-statistics-2026
- BCG — “AI at Work 2025.” n=10,600+ workers, 11 countries. Only 5% achieving substantial AI returns. Independent survey. High credibility. https://www.bcg.com/publications/2025/ai-at-work-momentum-builds-but-gaps-remain
- WalkMe/Propeller Insights — Shadow AI survey. n=1,000 U.S. workers, July 2025. ±3% margin of error. 78% unapproved AI tool usage. Independent polling. High credibility. https://news.sap.com/2025/08/new-walkme-survey-shadow-ai-rampant-training-gaps-undermine-roi/
- LayerX — Enterprise AI & SaaS Data Security Report. October 2025. 77% paste data through personal accounts. Browser telemetry. Independent security vendor. High credibility. https://layerxsecurity.com/blog/layerxs-enterprise-genai-security-report-2025-exposing-hidden-ai-security-blind-spots/
- DX Research/Atlan — Year 1 TCO analysis. License fees = 10-20% of total AI deployment cost. 2.5x Year 1 multiplier. Independent. High credibility.
- Faros AI — 10,000+ developers, 1,255 teams. Review overhead and PR volume data. Practitioner data. Moderate credibility. https://www.faros.ai
- Fair Observer — “Why 95% of Enterprise AI Projects Fail.” 2025. 56% lost sponsorship, 61% treated as IT. Aggregating MIT, HBR, and consulting data. Moderate credibility.
- IBM — Cost of a Data Breach 2025. n=604 organizations, 17 countries. 63% lacked AI governance policies. Gold standard for breach cost data. https://www.ibm.com/reports/data-breach
- BCG — “From Potential to Profit.” n=1,250+, September 2025. 10-20-70 budget framework. Consulting survey. Moderate-high credibility. https://www.bcg.com/publications/2025/closing-the-ai-impact-gap
- McKinsey — “The State of AI in 2025.” n=1,993. June-July 2025. 6% with measurable EBIT impact. Independent survey. High credibility. https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai
Brandon Sneider | brandon@brandonsneider.com March 2026