Pitch Deck Prompt: Build the Fundraising Narrative Arc

Founders don't lose rounds because the slides were ugly. They lose them because the story didn't hold — the problem felt small, the why-now was missing, the ask landed before the investor believed the rest. A pitch deck prompt that builds the narrative arc first, before anyone opens a design tool, fixes the part that actually decides the meeting.

Search this and you get two kinds of pages. Template galleries that hand you ten slide names and a "10/20/30 rule." And storytelling essays that explain narrative arcs with case studies and end on a "we'll design your deck" service pitch. Both are advice. Neither hands you a prompt that takes your facts and returns your arc.

This post is the prompt. The arc it enforces, the variables it reads, and the refusal rule that stops the model from inventing the traction you wish you had.

Why a deck is a story, not a slide list

The classic arc is five acts: make the investor feel the problem, show the insight that cracks it, prove why now is the moment, back it with traction, then ask. Every strong deck follows roughly that shape because it mirrors how people actually get convinced — pain, then a reason to believe, then proof, then the request.

A slide template skips the hard part. It tells you to have a "problem slide" without telling you whether your problem connects to your insight, or whether your why-now is real or decorative. The thinking — does this story cohere? — is the work. A template can't do it. A prompt with the arc baked into its output contract can at least force the question.

The opinion worth holding: most decks fail at why-now, and the prompt should interrogate it hardest. Founders nail the problem and the solution because they live them. The why-now — what changed in the world that makes this possible or urgent this year — is where the story goes soft. A good pitch deck prompt refuses to accept a vague why-now and pushes back until there's a real one.

What you can do with a pitch deck prompt

• Turn raw company facts into a structured problem, insight, why-now, traction, ask arc
• Pressure-test the why-now until it names a specific shift, not "AI is hot right now"
• Draft the problem statement so it makes the reader feel the pain before the solution appears
• Frame the ask around milestones ("launch the second product") instead of a spend breakdown
• Generate two or three narrative angles from the same facts, then pick the sharpest
• Produce the talk-track that sits behind each slide, which is what you actually present

Anatomy of the prompt

Three parts, as always: a tight role, the company facts as context, and an output contract locked to the five-act arc. The facts go in. The story spine comes out — and only from facts you supplied.

Variables
  {{company_facts}}  → what you do, for whom, the traction numbers, the team
  {{market_context}} → the shift that makes now the moment (the why-now raw material)
  {{raise_details}}  → amount, stage, the milestones the money buys
  {{audience}}       → optional: seed angels vs. Series A institutional

Prompt
  Role: a partner at a seed fund coaching a founder on narrative.
  Task: build the fundraising arc from the facts below. Every claim
        traces to a supplied fact. If a section has no supporting fact,
        write "needs input: <what's missing>" — never invent a number.
        Push back on a weak why-now.

Output contract
  ## Problem    (make the reader feel it; cite the fact that sizes it)
  ## Insight    (the non-obvious thing you know that others don't)
  ## Why now    (the specific shift; flag if thin)
  ## Traction   (only real numbers from {{company_facts}})
  ## The ask    (amount · milestones it buys · not a spend table)

That contract is why this beats a slide-name list. It enforces the logic between sections, not just their presence.

Step-by-step usage

1. Gather the facts

Dump everything into {{company_facts}}: what you do, who pays, real traction numbers, the team's unfair advantage. Be honest about the numbers — the prompt's whole value is that it won't invent the ones you leave out. Put the market shift into {{market_context}} and the raise into {{raise_details}}.

2. Set the refusal boundary

This line keeps the deck credible:

Every claim traces to a fact in {{company_facts}} or {{market_context}}.
If a section has no supporting fact, write "needs input: <what's missing>".
Never fabricate traction, market size, or customer counts.

The "needs input" output is a feature. It tells you exactly which part of your story has no evidence yet — which is the part an investor will poke.

3. Run it model-aware

Claude holds the five-act arc coherently across a long brief and is good at the why-now pushback when you ask for it explicitly. GPT-4o writes punchier problem statements but needs the arc restated on the final line of the prompt, or it reverts to a flat list of sections. Gemini is the most likely to soften the "needs input" flags into confident-sounding filler, so keep the no-fabrication rule loud for it.

4. Post-process

Read the arc aloud. If the insight doesn't follow from the problem, or the ask arrives before you believe the traction, the story has a seam. Regenerate that section with a sharper fact rather than rewriting by hand — the model will reconnect it better than a patch will.

5. Iterate

Ask for two alternative why-now framings from the same {{market_context}}. The first is usually the obvious one; the second or third is often the angle that makes the round feel inevitable.

Prompt-craft patterns that hold

Fact-or-flag. Every claim cites a supplied fact, or the section flags itself as missing input. This is the rule that separates an honest narrative from a confident fiction.

Rule: claim → supporting fact. No fact → "needs input". Never invent.

Feel before solve. Force the problem section to land emotionally before any solution appears. Investors decide whether they care in the first ninety seconds.

Problem section: make the reader feel the cost of the status quo
before naming the product. Cite the number that sizes the pain.

Milestone asks, not spend tables. The ask should name what the money achieves, not how it's divided. The fundraising narrative playbook frames the ask this way by default, because "18 months of runway to launch product two" beats a pie chart every time.

Variables you'll set

Getting started

1. Dump honest company facts and traction numbers into {{company_facts}}.
2. Put the market shift into {{market_context}} and the raise into {{raise_details}}.
3. Paste the role, the fact-or-flag refusal rule, and the five-act contract last.
4. Run it on Claude, which holds the arc and pushes back on a weak why-now.
5. Read the arc aloud and find the seam where the logic skips.
6. Regenerate that section with a sharper fact instead of hand-patching.
7. For the arc, the refusal rule, and the milestone-ask format already wired, use the Fundraising Narrative playbook.

The narrative is the spine, but a raise touches other documents. Once the story holds, the product roadmap prioritization prompt turns the milestones into a credible plan, and the technical design doc prompt backs the "how we build it" question when a technical investor asks.

A closing caution. The prompt structures your story; it can't make a thin story strong. If the why-now keeps coming back as "needs input," that's not a prompt failure — it's the round telling you the timing isn't argued yet. Treat the output as a sharp draft, pin the model version when a deck matters, and never let a fabricated number reach a slide. For the board-facing version after you raise, the Board Pack Production pack takes the same facts into a quarterly update, and the PRD prompt turns the roadmap promise into a spec.