How-To Guide · 2026

How to Write a Pitch Deck with Claude - narrative first, slides second.

The mistake founders make is trying to design slides before they've locked the narrative. Claude is excellent at narrative. Use it to build your story arc, write each slide's headline and supporting copy, and pressure-test your logic - before you touch Canva or Keynote.

The Short Version

Feed Claude your company facts, your target investor thesis, and the 10-slide structure. Ask it to write the narrative arc first (2 paragraphs), then generate each slide's headline and 3 supporting bullets. Your editing pass: replace generic statements with your actual metrics and make the "Why now" and "Why us" slides viscerally specific. Total time: 2–3 hours for a deck that would have taken 6–10.

What inputs Claude needs

Before writing a single slide, Claude needs to understand your company, your raise, and who you're pitching. Give it:

Company description in 3 sentences - what you do, who for, what outcome
The problem you're solving - include real evidence the problem exists (market data, customer quotes)
Your solution - mechanism of how it works, what makes it different
Traction - ARR, MRR, customer count, growth rate, retention, whatever you have
Market size and framing - how you define TAM/SAM, with sources
Business model - how you make money, unit economics if available
Team - relevant background for each founder, why you're the people to build this
The ask - raise amount, use of funds, valuation context if you have it
Target investor - stage, thesis, portfolio (so Claude can calibrate the language)

The pitch deck system prompt

You are an investor pitch deck writer who has helped 50+ founders raise seed through Series B rounds. Your job is to write the narrative and slide copy for a pitch deck. STRUCTURE (10 slides): 1. Cover - company name, one-liner, founder names 2. Problem - the pain, with evidence it's real and widespread 3. Solution - what you built and the key insight behind it 4. Why Now - the timing argument (technology shift, regulatory change, market inflection) 5. Traction - metrics that prove people want this 6. Market - TAM/SAM with a bottoms-up argument 7. Business Model - how you make money and the path to strong unit economics 8. Competition - honest positioning, why you win 9. Team - specific credentials that make you uniquely qualified 10. The Ask - amount, use of funds, what you'll hit by next raise PROCESS: First, write a 3-paragraph narrative arc (the story the deck tells from problem to ask). Then, for each slide, write: (a) the slide headline (8 words or fewer), (b) 3 supporting bullets, (c) any key stat or quote to anchor the slide. RULES: - Never write "we are disrupting" or "game-changing" or "world-class" - The problem slide must make an investor feel the pain, not just understand it - Every metric must be clearly defined (ARR not "revenue", NRR not "retention") - The team slide must explain credentials, not just list titles COMPANY BRIEF: [Paste all inputs here]

Workflow: brief → prompt → draft → edit

Step 1 - Brief (30–45 min): This is the thinking work. Write out each input listed above in a Google Doc or notes file. Don't skip the "Why Now" - it's the slide most founders underinvest in and the one good investors probe hardest.

Step 2 - Narrative first (10 min): Before asking Claude for slides, ask it to write the 2–3 paragraph narrative arc. If the narrative doesn't make intuitive sense, the slides won't either. Fix the logic at this stage, not after you've designed 10 slides.

Step 3 - Slide copy generation (10–15 min): Paste the system prompt with your full brief. Ask Claude to generate all 10 slides in the format: slide title, headline, 3 bullets, key stat. Review the full output before requesting any changes.

Step 4 - Editing pass (45–60 min): Replace every generic statement with a specific one. "Strong growth" → "3.2x ARR growth in 12 months." "Experienced team" → "Sarah was employee #4 at [Acquired Company], led the data infrastructure that supported $200M ARR." These specifics are what Claude can't generate - they're what you have to bring.

Common mistakes

Letting Claude invent market size numbers. Claude will write plausible-looking TAM figures. Every market size number in your deck must come from a real source - Gartner, IDC, your own bottoms-up calculation. Replace Claude's numbers with yours.
Keeping Claude's generic "Why Us" copy. This is the most important slide and the one Claude does worst. Rewrite it completely with specific, verifiable reasons your team is uniquely positioned.
Not pressure-testing the logic. After the draft, ask Claude: "What objections would a skeptical seed investor have to this deck?" This surfaces gaps before a real investor does.
Skipping the narrative-first step. If your narrative arc doesn't flow, the individual slides won't save it. Check the story before fixing copy.

Time savings

First-time pitch decks typically take founders 15–40 hours to write (including rewrites after feedback). Using this workflow, founders report getting to a review-ready draft in 3–5 hours. Revision cycles are also faster because the narrative structure is already sound - you're wordsmithing, not restructuring.

The bigger gain is quality: Claude holds the structure accountable. It won't let you accidentally skip the "Why Now" slide or write a problem slide that's really a solution slide.

What Claude can't do

Generate your actual metrics - those come from your business
Validate your market sizing methodology - verify against real data sources
Replace investor feedback - you still need humans who've seen 500+ decks to pressure-test yours
Design the deck - Claude writes the words; you need a designer or Figma template for the visuals