Pitch Deck Prompt for Founders

What this prompt does

A pitch deck prompt turns the facts only a founder has, like what your first paying customer said and what they paid, into a slide-by-slide outline for a seed round — with every number traced back to the line you typed. You supply the evidence. The model supplies the order, the compression, and a list of what's still missing. It does not supply a market size, a growth rate, or a raise amount, because the version of you sitting across from an investor has to defend every one of those out loud.

That's the difference between this and pasting "write me a pitch deck" into a chat window. Ask a model to write the deck from scratch and it hands back the one every investor has already sat through — a problem slide full of adjectives and a market figure nobody can source. Ask it to compress what you already know and it earns its place, because shortening a sentence you wrote is something a model does well, and writing the sentence you haven't is something it fakes.

The client presentation prompt is the nearest neighbour, and the two decks are built for opposite rooms. That one structures a consultant's delivery deck for a client who already bought; this one structures a founder's fundraising story for an investor who hasn't. The elevator pitch prompt covers the sixty seconds you say out loud. This page covers the document that follows.

The prompt

Build a seed-stage pitch deck outline from the inputs below. Compress what I give you. Do not add claims I didn't make.

**What we do, in one sentence:** [ONE PLAIN SENTENCE, NO ADJECTIVES. If you have two versions, paste both.]
**The problem, in a customer's words:** [PASTE A REAL QUOTE, VERBATIM, and who said it — role and company type, no names]
**What we built and what it does today:** [WHAT EXISTS NOW — not the roadmap]
**Traction, with dates:** [REAL NUMBERS — signups, paying customers, revenue, retention, the price someone paid, and the month each figure is from]
**Why now:** [WHAT CHANGED in the last two or three years that makes this possible or urgent]
**How we make money:** [WHO PAYS, HOW MUCH, HOW OFTEN]
**How people solve this today:** [THE ALTERNATIVES — competitors, spreadsheets, doing nothing]
**Team:** [FOR EACH FOUNDER, the one or two facts that matter for this company specifically]
**The ask:** [YOUR NUMBER, what it buys, and the milestone it reaches — e.g. "$600,000 to reach 1,000 paying studios in 18 months"]
**Market figure, if I have one:** [A NUMBER + ITS SOURCE + A LINK — or write "none yet"]
**What isn't proven yet:** [BE HONEST]

Output:
1. Three versions of my one-sentence description, each shorter than mine. Don't add words I didn't use.
2. A 10-slide outline in this order: purpose, problem, solution, why now, traction, business model, market, competition, team, the ask. For each slide give a headline of 12 words or fewer, the on-slide text in 30 words or fewer, and a "from:" line naming which input each number came from.
3. A "go and get" list: the three inputs that would most strengthen this deck, stated as the specific thing to collect — not as slide content.

Rules:
- Every number in the outline must appear in my inputs. If a slide needs a figure I didn't give, write [VERIFY: what's needed] and leave it.
- If I wrote "none yet" for the market figure, the market slide states the customer and the segment only. Do not estimate a market size.
- Repeat my ask exactly. Never suggest a different amount, a valuation, or deal terms.
- No adjectives from this list anywhere: revolutionary, disruptive, innovative, passionate, seamless, unique, world-class, cutting-edge.
- If an input is thin, the slide stays thin. Say so in the "go and get" list instead of padding it.

How to use it

1. Write the one-sentence description before you open the chat, badly if necessary. Sequoia's pitch guide puts it first for a reason: define the company in a single declarative sentence, and expect it to be harder than it looks. The model's job is to shorten yours. It can't shorten a blank.
2. Pull the traction numbers from the source, with dates. Pull them from the Stripe dashboard or the bank export, not from memory. "About 200 customers" becomes "214 paying studios as of August 31" and the deck stops sounding like a guess. Undated numbers are the first thing a careful investor asks about.
3. Paste the customer quote verbatim, typos included. Your problem slide is the only one an investor can't fact-check against a spreadsheet, and a real sentence from a real person carries more weight than anything the model can phrase. If you don't have a quote, that's the first item on your "go and get" list.
4. Read the "from:" lines before you read the slides. Every number should point at an input. One that points nowhere is an invention, and it goes.
5. Treat the "go and get" list as the week's to-do. A thin slide is an evidence problem, and the model just told you which evidence.

Example output (partial)

Inputs: studio management software for independent pottery studios; customer quote from a studio owner, "I spend Sunday nights matching kiln slots to a spreadsheet and I still double-book"; 214 paying studios at $49/month as of August 31, 2026, up from 96 in April; month-over-month growth averaging 22% over four months; monthly churn 2.3%; 11 of the first 20 customers came from one thread in a ceramics forum; ask: $600,000 to reach 1,000 paying studios in 18 months; market figure: none yet.

One-sentence options

1. Kiln scheduling and memberships for independent pottery studios.
2. Scheduling software for pottery studios.
3. Studio software for potters.

Slide 2 — Problem Headline: Studio owners lose Sunday nights to a kiln spreadsheet that still double-books. On slide: "I spend Sunday nights matching kiln slots to a spreadsheet and I still double-book." — owner, 40-member studio from: customer quote input

Slide 5 — Traction Headline: 214 studios pay $49 a month, up from 96 in April. On slide: 214 paying studios (Aug 31, 2026) · 22% average monthly growth over four months · 2.3% monthly churn from: traction input

Slide 7 — Market Headline: Independent pottery studios that run on memberships and kiln time. On slide: [VERIFY: number of independent studios in your first country, with a source] from: market figure input ("none yet")

Slide 10 — The ask Headline: $600,000 to reach 1,000 paying studios in 18 months. On slide: $600,000 · 1,000 studios · 18 months from: ask input

Go and get

1. A sourced count of independent pottery studios in your first market — the market slide is empty without it.
2. Revenue retention by cohort, not just churn. Four months of data will support a cohort chart.
3. A second customer quote about the membership side. Every quote you have is about kilns.

One slide stayed empty in that example, and that's the prompt working. An empty slide gets filled with a number you can defend — a filled one gets defended with a number you can't.

Variations

For a Series A

"This is a Series A deck. Add two slides after traction: unit economics from the figures I give you, and a use-of-funds breakdown of my ask in three lines. Same rules — no figure I didn't supply."

For the version that gets emailed

Add "this deck will be read without me in the room, so each on-slide text must make sense with no narration, and the ask goes on slide 2 as well as slide 10." An emailed deck can't rely on you to bridge the gaps, and YC's seed-deck template notes that when founders stretch a thin story, investors get lost. Clarity beats completeness in a deck that has to speak for itself.

Pre-revenue

Replace the traction input with "Evidence so far: [waitlist size, letters of intent, pilot results, interviews completed — with dates]" and add "there is no revenue yet; do not phrase any evidence as if it were." Run the case study prompt on your best pilot; that story is what carries this slide.

The one-pager that goes with it

Run the executive summary prompt on the finished outline, not the other way round. Summarizing a finished deck takes twenty minutes. Writing the summary first means building the whole argument twice — once without the slides to lean on.

Common pitfalls

• Don't let the model size the market for you. Name a segment and it will produce a figure, complete with a plausible-looking source, and the first follow-up question in the room will find out the source doesn't exist. The competitive analysis prompt is the better use of that slide's time: who solves this today and why they haven't solved it well.
• Don't ask it what to raise. The ask is yours — worked out from your burn and your milestones, with people who know your numbers. The prompt repeats it; it never proposes one, and it never touches valuation or terms.
• Don't feed it abstractions. "Strong early traction" gets you a slide that says "strong early traction." 214 studios at $49 a month gets you a slide an investor can do arithmetic on.
• Don't pad thin slides. A model asked to fill ten slides will fill ten slides, and the padding reads exactly like padding. The prompt's "if an input is thin, the slide stays thin" rule is the one you'll be most tempted to delete. Leave it in.
• Don't paste the business plan prompt's output as an input. The plan is for your thinking. The deck is for a room with twenty minutes. They compress differently.

Who uses this prompt

• First-time founders: the deck stops being a blank page and becomes a checklist of evidence to collect, which is the more useful thing for it to be at this stage. More in prompts for entrepreneurs.
• Founders on their second raise: the Series A variation, run against real cohort data, produces a draft in an afternoon that used to take a week of slide fiddling.
• Accelerator cohorts: every founder in a batch runs the same prompt on their own inputs, and the mentors review evidence instead of formatting.
• Solo founders without a designer: the outline is the hard part. Getting it onto slides is an evening.

ChatGPT, Claude, and Gemini all run this prompt unmodified. It leans on two ideas from Anthropic's prompting best practices: treat the model as a capable new hire who has none of your context, so give it the context, and explain the reason behind a constraint rather than just stating it. That's why the prompt says the deck will be defended out loud, not just "don't invent numbers." Eleven of this example's first twenty customers came from one forum thread. That fact never made a slide, and it's the first thing a good investor would ask about.