The 30-Day AI Champion Bootcamp · Lesson 4 of 4
Week 4 — The Pitch
> 30 min/day · 6 days · A 5-slide deck and a meeting on the calendar by Saturday
You have a problem brief. You have a working prototype. You have a real user's reaction on video. This last week is about packaging all of that into something leadership can say yes to in 20 minutes.
Most AI pitches fail for one of three reasons. They're too technical ("we used embedding vectors and RAG"), too vague ("AI will transform our business"), or ask for too much ("we need $500K and 4 engineers"). Yours is going to be the opposite of all three.
Not 15. Not 8. Five. Here's what's on each one.
Slide 1 — The Cost of Doing Nothing
The first slide is not "about this presentation" or your name or the agenda. It's the problem, in dollars, with a real person named on it.
``` Joe in Accounts Payable spends 3 hours every Monday chasing invoice approvals.
52 weeks × 3 hours × $32/hr loaded = $5,000/year.
And Joe hates Mondays. ```
That's the whole slide. One photo of the spreadsheet chaos in the background if you must, but three lines of text is enough. Leadership wakes up.
Slide 2 — What I Built
No theory. No architecture diagram. A screenshot of your prototype and a screenshot of the user using it.
If you can embed the 5-minute Loom from Week 3 here, do it. If the room is the kind where people don't actually watch videos, use two screenshots side by side: "Before" (their manual spreadsheet / 17-email thread / PDF search) and "After" (your chatbot with a clean answer).
One sentence of caption: *"I built this over four weeks on my own time. [User's name] tried it last week."*
Slide 3 — What It Does
Bullets. Not paragraphs. The narrower, the better.
```
- Reads the invoice email
- Extracts the approver, amount, and due date
- Drafts a tactful reminder email
- You still click Send (for now)
```
That last line is a gift to leadership. It says: *I'm not asking you to trust AI to do the whole thing. I'm asking for permission to keep going.*
Slide 4 — The Math That Scales
The ROI case. Your worksheet from Week 1 grew up.
``` Today: 1 user (Joe), 3 hrs/week = $5,000/year saved
If we roll to the AP team (4 people): $20,000/year saved
If we use the same pattern on the sales follow-up process that Jenna does manually: another $18,000/year
v1 investment: 4 hours of my time + ~$30/month tool cost ```
The trick here is not overselling the present — it's showing that the *pattern* scales. One prototype, proven. Three more within reach. All with the same tooling they already pay for.
Slide 5 — The Ask
One specific thing. Ideally three things in order of aggressiveness, so they can say yes to the easy one if not the big one.
``` What I'm asking for today:
Minimum: 4 hours of IT time to move this from my laptop to a test tenant so the whole AP team can try it.
Ideal: $5K/quarter AI tools budget + 20% of my time as the "AI lead" for the next 90 days.
Moonshot: A real AI Champion role + $30K annual AI budget + a cross-functional squad to ship one wedge a quarter. ```
Now they have *something to say yes to.* Not an all-or-nothing ask. Not a vague "invest in AI."
Day 1 — Draft (30 min)
Open the 5-Slide Pitch Deck Template. Fill in the rough copy for each slide using the structure above. Don't worry about design. Don't pick fonts. Raw words first.
If you're stuck on the ask slide, that's the signal you need to go talk to your manager before this meeting. Not to pitch — to ask *"if I could get one thing from leadership for a project like this, what's the realistic shape of it here?"*
Day 2 — The numbers pass (30 min)
Go back to your ROI worksheet. Make sure every number in your deck is traceable to a specific assumption.
- Hours/week: from your user interview
- Hourly cost: your company's loaded rate, or $32/hr as a conservative default
- Weeks per year: 50 (accounting for holidays)
- Tool cost: a quote from the actual product (Claude $20/mo, ChatGPT Team $30/user/mo, Copilot $30/user/mo, etc.)
If your CFO pressed you on any of these numbers, could you defend it? If not, rework it. Numbers you can't defend will tank the pitch no matter how good slide 1 is.
Day 3 — Design pass (30 min)
Don't waste more than 30 minutes on design. The content is 95% of the battle.
Minimum acceptable design bar:
- Consistent font (one sans-serif, pick any)
- One accent color (your company's brand color, or amber)
- No clip art, no stock photos, no "AI in a circuit board" hero images
- One idea per slide
- Numbers bigger than text
If you can make it look like something McKinsey would hand in, great. If you can only make it look like a clean Google Slides template, also great.
Day 4 — Rehearse (30 min)
Present the deck out loud, standing up if you can. Time yourself. It should be under 10 minutes with room for questions.
Common self-corrections:
- You spend too long on slide 2 explaining *how* you built it → cut that, no one cares yet
- You hedge on slide 5 → remove all "maybe," "if possible," "I was thinking." Just state the ask.
- You forget the user's name → don't. Their name is more persuasive than any technology word.
Record yourself on your phone. Watch it back at 1.5x. Painful but diagnostic.
Day 5 — Book the meeting (30 min)
Use the Leadership Meeting Email Script template.
The email has three parts:
Subject line: *"30-min request: a working AI prototype I want to show you"*
Body:
> Hi [Exec], > > Over the last month, I built a working prototype that automates the first step of [process]. [User] tested it last week and it saved them [N] minutes per [unit]. If we applied the same pattern across [team], the numbers get real — [$X]/year in reclaimed time. > > Could I grab 30 minutes with you in the next two weeks to walk through the prototype and a proposal? I've put together a 5-slide deck and I have the Loom of the live test. > > I'm also happy to loop in [manager] if that's easier. > > [Name]
Send it today. Not tomorrow. Today.
Day 6 — Followups + fallbacks (30 min)
If they haven't replied by Day 6 (about a week in exec time, which is normal):
- Forward the original email to your manager with: *"I sent this to [Exec] and haven't heard back. Could you help me land the meeting, or should I shift the ask?"*
- If the exec's assistant exists, cc them on the follow-up
- If they decline, ask your manager if you can present the same deck to them instead — middle management greenlight is a perfectly valid V1
Don't cold-follow-up three times. Two is the max. Then pivot.
You finish Week 4 with:
- A 5-slide deck + your Week 3 Loom, ready to go
- A calendar invite you sent to someone with budget authority
- A rehearsed, timed delivery
- An honest ROI story you can defend
After the meeting — three possible outcomes, each fine:
- Yes. Celebrate for 10 minutes, then go read *Chapter 2: The Math That Closes Deals* in the Applied AI Playbook for how to measure the rollout.
- Yes, but smaller than you asked. Take it. Small yes compounds.
- No, not right now. Ask "what would have to be true for this to be a yes in 6 months?" Write down the answer word-for-word. That's your next wedge. Take the prototype and the deck with you to interviews — it's the strongest portfolio piece you have.
- 5-Slide Pitch Deck Template — PowerPoint + Google Slides versions, already formatted
- ROI Worksheet (Advanced) — the defend-every-number version of your Week 1 math
- Leadership Meeting Email Script — the cold ask, the warm follow-up, the fallback
You spent 30 days going from "I use ChatGPT sometimes" to "I shipped a working prototype and pitched it to leadership." That's not an AI skill. That's the operator skill that 99% of your coworkers don't have.
Whether your specific pitch gets funded or not, you now know how to do this repeatably. The next wedge is always easier.
Congrats on finishing. Mark the course complete below to get your certificate — and then tell us what you pitched in the community feed. We're rooting for you.
— John
Previous: Week 3 — Test with One Real User · Back to the course