The Department Head's AI Pilot Pitch: How to Propose, Scope, and Defend a Team-Level Experiment

Executive Summary

• Over 52% of department-level AI initiatives operate without formal approval or oversight — which means they also operate without budget, air cover, or a path to production. The VP of Sales running an unsanctioned AI experiment has a tool. The VP who pitched it formally has an initiative with executive sponsorship, measurable metrics, and a 68% success rate versus 11% without that sponsorship (EY Autonomous AI Survey, March 2026; Pertama Partners, n=2,400+ initiatives, 2025-2026).
• Employees are moving faster than leadership realizes. McKinsey found that C-suite executives estimate 4% of employees use AI for a substantial portion of their work. The actual number is 13% — more than three times the estimate (McKinsey Superagency, n=3,613 employees + 238 C-suite, January 2025). The department head who formalizes what is already happening converts shadow usage into sanctioned value.
• The pitch that gets approved leads with the problem, not the technology. RAND Corporation identifies “technology over problem-solving” as a top-five root cause of AI project failure. The budget template on the next page forces the sequence: business pain first, proposed approach second, AI last.
• A 60-day bounded experiment with a named metric and a kill threshold is not a risk — it is a risk reduction. S&P Global found 42% of companies abandoned AI initiatives in 2025, most after 11 months and $4.2M in sunk costs. A 60-day pilot with a $15,000-$50,000 budget and a pre-defined stop condition bounds the exposure to a fraction of that (S&P Global, n=1,006, 2025).

Why This Pitch Exists — And Who It Is For

The COO’s pilot structure card tells you how to design a pilot. The CFO’s budget template tells you how to request funding. Neither solves the problem facing the VP of Marketing, the Head of Operations, or the Director of Customer Success who sees a clear opportunity and needs to sell it internally — to a CIO who controls tool procurement, a CFO who controls budget, and a CEO who controls strategic priority.

That person does not speak CIO. They should not have to.

McKinsey’s Superagency report (n=3,613, January 2025) documents the gap: 47% of employees believe AI could handle 30% or more of their work within a year. Only 20% of executives agree. The department head sits in the middle of that perception gap — close enough to the work to see the opportunity, senior enough to make the case.

The pitch below takes 45 minutes to complete. It produces a one-page document that answers every question the CIO, CFO, and CEO will ask — in language none of them need to translate.

The One-Page Pilot Pitch

Section 1: The Problem You Are Solving (Not the Tool You Want)

This is the section most department heads get wrong. “I want to try AI for lead scoring” is a technology request. “The sales team spends 11 hours per week manually qualifying leads, resulting in a 23% false-positive rate and $180,000 in annual wasted effort on dead-end prospects” is a business problem.

RAND Corporation’s analysis of AI failure modes ranks “technology over problem-solving” among the top five root causes. The department head who leads with a tool name has already lost the pitch. The one who leads with a dollar figure has the CFO’s attention.

Fill in: “My team currently spends _____ hours per [week/month] on [specific task]. This costs approximately $_____ per year in [labor/errors/rework/delays]. The current process produces [volume] outputs at [quality metric]. The specific pain: [one sentence describing what breaks, slows, or costs money].”

Section 2: What You Propose — In Plain Language

No architecture diagrams. No vendor feature comparisons. One paragraph that a non-technical executive reads in 30 seconds.

OpenAI’s enterprise guide (based on 300+ implementation case studies and 2M+ enterprise users, April 2025) categorizes most AI use cases into six patterns: content creation, research and synthesis, coding, data analysis, ideation, and automation. The department head does not need to understand the technology. They need to identify which pattern matches their problem.

Fill in: “I propose a 60-day experiment where [number] team members use [tool or approach] to [specific action] on [specific workflow]. The experiment tests whether AI can [reduce time / improve accuracy / increase throughput] on this one task. If it works, the team reclaims approximately [hours/week] for [higher-value activity].”

Section 3: What Success Looks Like — One Number

Not three numbers. Not a dashboard. One metric that the CFO puts on a slide and the CEO reads in 10 seconds.

McKinsey (n=1,600+, November 2025) identifies KPI tracking as the single strongest predictor of bottom-line impact from AI — yet fewer than 20% of organizations track KPIs for their AI tools. The 80% that skip this step cannot prove their pilot worked even when it did.

The test: Can your CFO put this number on a slide? If the answer is no, simplify until it is yes.

Fill in: “Success means [metric] moves from [baseline] to [target] within 60 days. I will measure this by [specific method]. If the metric has not moved by Day 30, I will [adjust approach / escalate / recommend termination].”

Section 4: What It Costs — Honestly

The department head’s most common mistake: listing only the software license. CloudZero’s survey (n=500, March 2025) finds license fees represent 10-17% of total AI spend. The remaining 83-90% sits in implementation, training, data preparation, and the time your team spends learning a new tool instead of doing their current job.

At a 300-person company, a typical department-level AI pilot runs $15,000-$50,000 for 60 days when all costs are included. If your total is under $5,000, you have not accounted for team time. If it is over $100,000, this is not a pilot — it is a program, and it needs the full budget template.

Section 5: What Happens If It Fails — And How You Will Know

This is the section that separates a pitch from a wish. Pertama Partners (n=2,400+, 2025-2026) found that projects with pre-defined success metrics achieve a 54% success rate versus 12% without them. The kill threshold is not pessimism — it is the discipline that earns trust.

Fill in: “If [metric] has not improved by at least [minimum threshold] at Day 30, I will [specific action: adjust the approach / reduce scope / recommend stopping]. If [metric] has not reached [target] by Day 60, the experiment ends. Total exposure if terminated at Day 30: $[amount]. Total exposure if terminated at Day 60: $[amount].”

Section 6: What You Need From Leadership

Name the specific approvals, not a vague “support.”

Fill in:

• From the CIO/IT: “[Tool] approved for [data type] use. Security review completed by [date].”
• From the CFO: “$[amount] approved for 60-day pilot from [budget line].”
• From the CEO/COO: “This initiative is a stated priority for Q[X]. My team has permission to spend [X hours/week] on this instead of [current activity].”

Three Questions You Will Get — And How to Answer Them

“Why can’t IT just handle this?”

IT handles tool procurement, security review, and technical configuration. They cannot answer the question this pitch answers: which business problem in your department justifies the investment? Cisco’s Working with AI program (Fortune, March 2026) reviewed 24 workflows and found an average of 30% of activities within each workflow could be augmented by AI. The CIO did not identify those workflows. The department heads who own the work did.

The answer: “IT is a critical partner for security and setup. This pitch defines the business case and success metric that IT needs to evaluate the tool against. Without this, IT is evaluating technology in a vacuum.”

“How do I know this won’t just become shelfware?”

The data supports the concern. S&P Global found 42% of companies abandoned the majority of AI initiatives in 2025. Worklytics benchmarks show generic AI training achieves 23% sustained adoption, while role-specific training reaches 67% (Worklytics, 2025).

The answer: “The pilot is scoped to one workflow with one metric. The team members are selected because they do this task daily and have expressed interest. Adoption is measured weekly. If adoption drops below [threshold] by Day 30, the pilot adapts or stops — it does not drift.”

“What about the AI tools we already pay for?”

This is the CFO’s sharpest question. Most mid-market companies already pay for AI features embedded in Microsoft 365, Salesforce, Google Workspace, and Zoom that nobody has evaluated or activated.

The answer: “I checked. [Existing tool] includes [AI feature], which [does/does not] address this specific workflow because [reason]. This pilot [uses that existing capability at no additional license cost / requires [specific tool] because existing tools do not cover [specific gap]].”

What the Data Shows About Department-Level Pilots That Work

The evidence on which departments capture AI value fastest is consistent across multiple surveys.

Worklytics’ 2025 benchmarks show median AI adoption rates vary dramatically by function: technology and engineering at 65-75%, sales and marketing at 55-70%, customer success at 60-75%, HR at 45-60%, and finance and operations at 40-55%. The departments with the highest adoption rates share a common trait: their work involves high-volume, repeatable tasks with measurable outputs.

BCG (n=10,635, June 2025) found that frontline employees have hit a “silicon ceiling” — only half regularly use AI tools. The bottleneck is not technology access. It is the absence of a structured path from individual experimentation to team-level practice. The department head who creates that path — with a formal pitch, a named metric, and a 60-day timeline — bridges the gap that informal adoption cannot.

The California Management Review’s evidence-based framework (November 2025) documents the “missing middle” problem: organizations can demonstrate AI’s technical feasibility in pilots but cannot translate that into scaled impact. Fewer than 40% of automation initiatives deliver measurable value (Deloitte CFO Survey, 2025). Only 30% of AI pilots transition to scaled impact (McKinsey Global AI Survey, 2025). The pilots that do scale share one characteristic: they started with a business problem owned by the person closest to the work — not a technology initiative owned by IT.

Key Data Points

What This Means for Your Organization

The gap between “everyone knows AI matters” and “someone actually proposed a structured experiment” is where most mid-market companies stall. The CIO is waiting for a business case. The CFO is waiting for a bounded request. The CEO is waiting for someone to take ownership. The department head who fills in this one-page pitch is not requesting permission to play with technology — they are demonstrating the operational discipline that predicts whether AI delivers value or becomes another line item nobody can justify at renewal.

The pitch takes 45 minutes. The 60-day pilot costs a fraction of the $4.2M average that companies spend before abandoning unstructured AI initiatives. And the department head who runs it successfully has done something the CIO cannot do from a distance: proven that AI works on a specific workflow, with specific people, producing a specific result that the CFO can put on a slide.

If completing this pitch raised questions about which workflow to select, how to define the right metric, or how to structure the conversation with the CIO and CFO — that is exactly the kind of question worth a 30-minute conversation before the first dollar moves. brandon@brandonsneider.com

Sources

1. McKinsey — “Superagency in the Workplace: Empowering People to Unlock AI’s Full Potential at Work.” n=3,613 employees + 238 C-suite leaders, surveyed October-November 2024, published January 2025. Source for 4% vs. 13% perception gap, 47% vs. 20% employee-executive readiness gap, 48% training priority. Independent survey. High credibility. https://www.mckinsey.com/capabilities/tech-and-ai/our-insights/superagency-in-the-workplace-empowering-people-to-unlock-ais-full-potential-at-work

2. Pertama Partners — “AI Project Failure Statistics 2026.” n=2,400+ enterprise AI initiatives, 2025-2026. Source for 54% vs. 12% metric success rate, 68% vs. 11% sponsorship impact. Independent consulting analysis aggregating RAND, MIT Sloan, McKinsey, and Deloitte data. High credibility. https://www.pertamapartners.com/insights/ai-project-failure-statistics-2026

3. S&P Global 451 Research — Voice of the Enterprise: AI & Machine Learning, Use Cases 2025. n=1,006, March 2025. Source for 42% initiative abandonment rate, $4.2M average sunk cost. Independent analyst. High credibility. https://www.spglobal.com/market-intelligence/en/news-insights/research/ai-experiences-rapid-adoption-but-with-mixed-outcomes-highlights-from-vote-ai-machine-learning

4. CloudZero — “State of AI Costs 2025.” n=500 U.S. software leaders, March 2025. Source for 10-17% license cost ratio. CloudZero sells cloud cost management (bias toward surfacing hidden costs). Medium credibility. https://www.cloudzero.com/state-of-ai-costs/

5. McKinsey — “The State of AI in 2025.” n=1,993 respondents across 105 countries, June-July 2025; n=1,600+ for KPI tracking finding, November 2025. Source for <20% KPI tracking, 30% pilot-to-scale rate. Independent survey. High credibility. https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai

6. RAND Corporation — AI project failure taxonomy identifying five root causes including “technology over problem-solving.” Independent, non-profit research organization. Very high credibility.

7. BCG — “AI at Work: Momentum Builds, but Gaps Remain.” n=10,635 across 11 countries, June 2025. Source for frontline “silicon ceiling,” 3.7x leadership support multiplier. Independent survey. High credibility. https://www.bcg.com/publications/2025/ai-at-work-momentum-builds-but-gaps-remain

8. Worklytics — “2025 Benchmarks: What ‘Good’ Employee AI Adoption Looks Like by Department and Industry.” 2025. Source for department adoption rates and 23% vs. 67% training effectiveness. Workplace analytics vendor (bias toward measurement solutions). Medium credibility. https://www.worklytics.co/resources/2025-employee-ai-adoption-benchmarks-by-department-industry

9. Fortune — “From Pilot Mania to Portfolio Discipline: How the Best Companies Are Escaping AI Purgatory.” March 19, 2026. Source for Cisco’s 24-workflow review, 30% augmentation finding. Independent journalism. High credibility. https://fortune.com/2026/03/19/from-pilot-mania-to-portfolio-discipline-ai-purgatory/

10. California Management Review — “Bridging the Gaps in AI Transformation: An Evidence-Based Framework for Scalable Adoption.” November 2025. Source for “missing middle” framework. Academic journal. Very high credibility. https://cmr.berkeley.edu/2025/11/bridging-the-gaps-in-ai-transformation-an-evidence-based-framework-for-scalable-adoption/

11. Deloitte — “State of AI in the Enterprise 2026.” n=3,235 senior leaders across 24 countries, August-September 2025. Source for <40% automation value delivery. Independent survey. High credibility. https://www.deloitte.com/us/en/what-we-do/capabilities/applied-artificial-intelligence/content/state-of-ai-in-the-enterprise.html

12. EY — “Autonomous AI Adoption Survey.” March 2026. Source for 52% of department initiatives lacking formal approval. Technology sector focus. Medium-high credibility. https://www.ey.com/en_us/newsroom/2026/03/ey-survey-autonomous-ai-adoption-surges-at-tech-companies-as-oversight-falls-behind

13. OpenAI — “Identifying and Scaling AI Use Cases.” Based on 300+ implementation case studies and 2M+ enterprise users, April 2025. Source for six use-case primitives. Vendor guide (bias toward AI adoption). Medium credibility — useful framework, vendor context. https://openai.com/business/guides-and-resources/identifying-and-scaling-ai-use-cases/

Brandon Sneider | brandon@brandonsneider.com March 2026