Pitch Deck
When to Use
- The founder is preparing a pitch deck for a fundraising round (pre-seed through Series A).
- The founder has an existing deck and wants structural or narrative feedback.
- The founder asks what slides to include or how to tell their story to investors.
Context Required
From startup-context: company one-liner, stage, product description, target customer, business model, traction metrics, team bios, fundraising history, and competitive landscape. If any of these are missing, prompt the founder before drafting slides.
From the user: target round size, target investor type (VC vs. angel), whether this is for a live pitch (fewer words, more visuals) or a send-ahead deck (more self-explanatory text).
Workflow
- Read startup context — Pull from
.agents/startup-context.md to populate slide content. Flag any gaps.
- Determine deck type — Ask if this is a live-pitch deck (visual-heavy, 30-40 words per slide max) or a send-ahead deck (can include more explanatory text, 60-80 words per slide).
- Draft the narrative arc — Before writing any slides, outline the story: what is the world like today (problem), what changes with your product (solution), why now, why this team, and what you need to get there.
- Write slide-by-slide content — Produce content for each of the 10-12 slides below. Each slide gets a title, key message, supporting points, and a suggested visual or data element.
- Review for investor lens — Check every slide against the question an investor would ask at that point. Flag weak spots.
- Produce final output — Deliver the deck outline as structured markdown. If the user wants a
.pptx, chain to the Anthropic pptx skill after content is finalized.
Output Format
Structured markdown with one H3 per slide. Each slide section contains:
- Title: The slide headline (concise, assertion-style, e.g., "Healthcare billing wastes $200B annually")
- Key message: The one thing the audience should remember from this slide
- Content: Bullet points or narrative text
- Visual suggestion: What chart, image, screenshot, or diagram belongs here
- Investor question this answers: The implicit question in the VC's mind
Frameworks & Best Practices
The 10-12 Slide Framework
| # |
Slide |
Purpose |
Common Mistakes |
| 1 |
Title / Hook |
Company name, one-liner, and a memorable hook stat or image |
Burying the one-liner; using a generic tagline |
| 2 |
Problem |
Make the pain visceral and specific to your ICP |
Being too abstract; citing a problem everyone already knows |
| 3 |
Solution |
Show what you built and how it eliminates the pain |
Feature-dumping; not connecting back to the problem |
| 4 |
Demo / Product |
Screenshot, GIF, or live product walkthrough |
Showing the admin panel instead of the user-facing magic |
| 5 |
Market Size |
TAM/SAM/SOM with a credible bottoms-up calculation |
Using only top-down "the market is $X trillion" numbers |
| 6 |
Business Model |
How you make money, unit economics, pricing |
Not showing actual or projected unit economics |
| 7 |
Traction |
The chart that goes up and to the right |
Vanity metrics; hiding the Y-axis; mixing timeframes |
| 8 |
Competition |
Why you win — positioning matrix or comparison table |
Claiming "no competitors"; using a 2x2 where you magically own the top-right |
| 9 |
Team |
Founders + key hires, relevant backgrounds |
Listing every employee; not explaining founder-market fit |
| 10 |
Go-to-Market |
How you acquire customers today and at scale |
Saying "we'll go viral" without a concrete channel strategy |
| 11 |
Financials / Ask |
How much you're raising, use of funds, key projections |
Not specifying what milestones the money unlocks |
| 12 |
Closing / Vision |
The big dream — where this goes in 5-10 years |
Being too conservative; forgetting contact info |
Narrative Arc Rules
- Slide 1-3: Establish tension. The audience should feel the problem before you show the answer.
- Slide 4-7: Build credibility. Prove you have a real product with real traction in a real market.
- Slide 8-10: Prove defensibility. Show you can win against alternatives and scale.
- Slide 11-12: Make the ask and paint the vision. End with ambition, not logistics.
Stage-Specific Guidance
- Pre-seed / Seed: Emphasize problem depth, founder-market fit, and early signals (waitlist, LOIs, design partners). Financial projections can be lighter.
- Series A: Emphasize repeatable go-to-market, unit economics, and a clear path from current traction to 3-5x growth. Investors expect real revenue data.
Principles
- Every slide title should be an assertion, not a label. "We grew 30% MoM for 6 months" beats "Traction".
- Remove any slide that does not advance the narrative. If you cannot articulate why a slide exists, cut it.
- The deck should be understandable by a partner who reads it at 11pm on their iPad without you presenting.
- Use no more than 3 fonts and 2 brand colors. Clutter kills credibility.
- Data beats adjectives. Replace "fast-growing" with "3x YoY".
Related Skills
investor-research — use before building the deck to tailor the narrative to specific investor interestsdata-room — prepare the backup materials investors will request after seeing the deckfundraising-email — write the outreach email that accompanies the deck
Examples
Example prompt: "Help me build a seed deck. We're raising $3M for an AI-powered contract review tool for mid-market legal teams. We have $40K MRR and 12 paying customers."
Good output snippet (Problem slide):
Slide 2: Problem
Title: Mid-market legal teams review 500+ contracts/month with 1990s tools
Key message: Contract review is the #1 time sink for legal teams at companies with 200-2,000 employees, costing $180K+/year in paralegal hours per company.
Content:
- 73% of mid-market companies still use manual redlining in Word
- Average contract review takes 45 minutes; 30 minutes is repetitive clause checking
- Errors in reviewed contracts cost companies 2-5% of deal value annually
Visual suggestion: Side-by-side showing a paralegal buried in Word docs vs. your product highlighting risks in seconds
Investor question this answers: Is this a real, painful, expensive problem?
Example prompt: "Review my existing deck — I keep getting passed on after the first meeting."
Good output approach: Read the deck, identify which of the 12 slides are missing or weak, check if the narrative arc has tension, and test every slide against the investor-question framework. Deliver specific rewrites, not vague advice.
1---2name: pitch-deck3description: When the user wants to create, review, or restructure a fundraising pitch deck for seed or Series A. Also activates when the user mentions "deck", "pitch", "investor presentation", or "slide structure".4---5
6# Pitch Deck
7
8## When to Use
9
10- The founder is preparing a pitch deck for a fundraising round (pre-seed through Series A).
11- The founder has an existing deck and wants structural or narrative feedback.
12- The founder asks what slides to include or how to tell their story to investors.
13
14## Context Required
15
16From `startup-context`: company one-liner, stage, product description, target customer, business model, traction metrics, team bios, fundraising history, and competitive landscape. If any of these are missing, prompt the founder before drafting slides.
17
18From the user: target round size, target investor type (VC vs. angel), whether this is for a live pitch (fewer words, more visuals) or a send-ahead deck (more self-explanatory text).
19
20## Workflow
21
221. **Read startup context** — Pull from `.agents/startup-context.md` to populate slide content. Flag any gaps.
232. **Determine deck type** — Ask if this is a live-pitch deck (visual-heavy, 30-40 words per slide max) or a send-ahead deck (can include more explanatory text, 60-80 words per slide).
243. **Draft the narrative arc** — Before writing any slides, outline the story: what is the world like today (problem), what changes with your product (solution), why now, why this team, and what you need to get there.
254. **Write slide-by-slide content** — Produce content for each of the 10-12 slides below. Each slide gets a title, key message, supporting points, and a suggested visual or data element.
265. **Review for investor lens** — Check every slide against the question an investor would ask at that point. Flag weak spots.
276. **Produce final output** — Deliver the deck outline as structured markdown. If the user wants a `.pptx`, chain to the Anthropic pptx skill after content is finalized.
28
29## Output Format
30
31Structured markdown with one H3 per slide. Each slide section contains:
32- **Title**: The slide headline (concise, assertion-style, e.g., "Healthcare billing wastes $200B annually")
33- **Key message**: The one thing the audience should remember from this slide
34- **Content**: Bullet points or narrative text
35- **Visual suggestion**: What chart, image, screenshot, or diagram belongs here
36- **Investor question this answers**: The implicit question in the VC's mind
37
38## Frameworks & Best Practices
39
40### The 10-12 Slide Framework
41
42| # | Slide | Purpose | Common Mistakes |
43|---|-------|---------|-----------------|
44| 1 | **Title / Hook** | Company name, one-liner, and a memorable hook stat or image | Burying the one-liner; using a generic tagline |
45| 2 | **Problem** | Make the pain visceral and specific to your ICP | Being too abstract; citing a problem everyone already knows |
46| 3 | **Solution** | Show what you built and how it eliminates the pain | Feature-dumping; not connecting back to the problem |
47| 4 | **Demo / Product** | Screenshot, GIF, or live product walkthrough | Showing the admin panel instead of the user-facing magic |
48| 5 | **Market Size** | TAM/SAM/SOM with a credible bottoms-up calculation | Using only top-down "the market is $X trillion" numbers |
49| 6 | **Business Model** | How you make money, unit economics, pricing | Not showing actual or projected unit economics |
50| 7 | **Traction** | The chart that goes up and to the right | Vanity metrics; hiding the Y-axis; mixing timeframes |
51| 8 | **Competition** | Why you win — positioning matrix or comparison table | Claiming "no competitors"; using a 2x2 where you magically own the top-right |
52| 9 | **Team** | Founders + key hires, relevant backgrounds | Listing every employee; not explaining founder-market fit |
53| 10 | **Go-to-Market** | How you acquire customers today and at scale | Saying "we'll go viral" without a concrete channel strategy |
54| 11 | **Financials / Ask** | How much you're raising, use of funds, key projections | Not specifying what milestones the money unlocks |
55| 12 | **Closing / Vision** | The big dream — where this goes in 5-10 years | Being too conservative; forgetting contact info |
56
57### Narrative Arc Rules
58
59- **Slide 1-3**: Establish tension. The audience should feel the problem before you show the answer.
60- **Slide 4-7**: Build credibility. Prove you have a real product with real traction in a real market.
61- **Slide 8-10**: Prove defensibility. Show you can win against alternatives and scale.
62- **Slide 11-12**: Make the ask and paint the vision. End with ambition, not logistics.
63
64### Stage-Specific Guidance
65
66- **Pre-seed / Seed**: Emphasize problem depth, founder-market fit, and early signals (waitlist, LOIs, design partners). Financial projections can be lighter.
67- **Series A**: Emphasize repeatable go-to-market, unit economics, and a clear path from current traction to 3-5x growth. Investors expect real revenue data.
68
69### Principles
70
71- Every slide title should be an assertion, not a label. "We grew 30% MoM for 6 months" beats "Traction".
72- Remove any slide that does not advance the narrative. If you cannot articulate why a slide exists, cut it.
73- The deck should be understandable by a partner who reads it at 11pm on their iPad without you presenting.
74- Use no more than 3 fonts and 2 brand colors. Clutter kills credibility.
75- Data beats adjectives. Replace "fast-growing" with "3x YoY".
76
77## Related Skills
78
79- `investor-research` — use before building the deck to tailor the narrative to specific investor interests
80- `data-room` — prepare the backup materials investors will request after seeing the deck
81- `fundraising-email` — write the outreach email that accompanies the deck
82
83## Examples
84
85**Example prompt**: "Help me build a seed deck. We're raising $3M for an AI-powered contract review tool for mid-market legal teams. We have $40K MRR and 12 paying customers."
86
87**Good output snippet** (Problem slide):
88
89> ### Slide 2: Problem
90> **Title**: Mid-market legal teams review 500+ contracts/month with 1990s tools
91> **Key message**: Contract review is the #1 time sink for legal teams at companies with 200-2,000 employees, costing $180K+/year in paralegal hours per company.
92> **Content**:
93> - 73% of mid-market companies still use manual redlining in Word
94> - Average contract review takes 45 minutes; 30 minutes is repetitive clause checking
95> - Errors in reviewed contracts cost companies 2-5% of deal value annually
96> **Visual suggestion**: Side-by-side showing a paralegal buried in Word docs vs. your product highlighting risks in seconds
97> **Investor question this answers**: Is this a real, painful, expensive problem?
98
99**Example prompt**: "Review my existing deck — I keep getting passed on after the first meeting."
100
101**Good output approach**: Read the deck, identify which of the 12 slides are missing or weak, check if the narrative arc has tension, and test every slide against the investor-question framework. Deliver specific rewrites, not vague advice.
yanacuti1121