Generate investment pitch decks from financial statements and market research automatically

: it's Tuesday morning, and a founder is in week three of creating their Series A pitch deck. They've extracted financial projections from a spreadsheet (manually, because copy-paste always breaks formatting), written a company overview from scratch, designed slides about market size using data they found in a PDF somewhere, and they're still nowhere near finished. Meanwhile, the investor meeting is in four days. This scenario repeats across hundreds of startups every month. Pitch decks are a necessary evil, but they're also a perfect candidate for automation. The core tasks are predictable: extract data from financial documents, analyse market research PDFs, structure that information into slides, and design them with visual hierarchy. None of this requires human creativity. What requires human judgment is strategy and narrative, which come after the deck infrastructure exists. The good news? You can automate the entire process from financial statement upload to ready-to-present slide deck in one orchestrated workflow. You don't need expensive design software or weeks of manual work. You need the right combination of tools connected properly.

The Automated Workflow

This workflow uses five tools orchestrated through n8n, which offers the best balance of flexibility and visual interface management for this particular task. The flow moves data sequentially from PDF parsing through infographic generation to final slide creation.

Step 1: Upload and Extract Financial Data

Start by setting up a webhook in n8n that accepts PDF uploads. When a founder uploads their financial statement, the webhook triggers immediately.

CopyPOST /webhook/pitch-deck-upload
Content-Type: application/json { "filename": "Q1_2026_financials.pdf", "file_url": "https://storage.example.com/uploads/statement.pdf"
}

From there, send the PDF to Chat With PDF by Copilot.us using its API endpoint. This tool can simultaneously analyse multiple PDFs and extract structured data through conversational queries.

CopyPOST https://api.copilot.us/chat
Authorization: Bearer YOUR_API_KEY
Content-Type: application/json { "documents": ["Q1_2026_financials.pdf", "market_research.pdf"], "query": "Extract the following in JSON format: revenue, expenses, burn rate, runway months, gross margin, customer acquisition cost, and lifetime value. Return only valid JSON.", "response_format": "json"
}

Chat With PDF returns structured JSON with the financial metrics. Store this in n8n's workflow variables for the next step.

Step 2: Analyse Market Research and Create Context

In parallel, process any market research PDFs the founder has uploaded. Create a prompt for Claude Opus 4.6 that analyses market size, competitive positioning, and growth opportunity based on the research documents.

CopyPOST https://api.anthropic.com/v1/messages
Authorization: Bearer YOUR_API_KEY
Content-Type: application/json { "model": "claude-opus-4.6", "max_tokens": 1500, "messages": [ { "role": "user", "content": "Based on this market research document, extract and summarise: (1) Total addressable market size, (2) Market growth rate, (3) Key competitors and their market share, (4) Customer problem identified, (5) Recommended positioning. Return as JSON with these keys: tam, growth_rate, competitors, pain_points, positioning. Be specific with numbers." } ]
}

Claude generates market context that's both data-driven and narrative-ready. This becomes the foundation for your deck's market slides.

Step 3: Generate Infographic Assets

Now you have structured financial and market data. Convert this into visual assets using Text2Infographic. Create separate requests for revenue trends, market opportunity, and customer metrics.

CopyPOST https://api.text2infographic.com/generate
Authorization: Bearer YOUR_API_KEY
Content-Type: application/json { "title": "Revenue Growth and Margin Expansion", "data": { "revenue": [1200000, 1850000, 2650000], "gross_margin": [0.48, 0.55, 0.62] }, "chart_type": "combined_bar_line", "color_scheme": "professional_blue", "output_format": "png"
}

Text2Infographic returns high-quality PNG files that look polished without requiring a designer. Store the URLs returned.

Step 4: Generate UI Components with v0

This is where most people fail: they try to hand-build slides in PowerPoint. Instead, use v0 to generate React components for each slide type. For each slide category (title, financial summary, market analysis, team), create a v0 request that generates the HTML/CSS.

CopyPOST https://api.v0.dev/generate
Authorization: Bearer YOUR_API_KEY
Content-Type: application/json { "prompt": "Create a professional investor pitch deck slide component in React. Show financial metrics: Revenue $2.65M (YoY growth 121%), Burn rate $85k/month, Runway 24 months, Gross margin 62%. Use a dark background with white text, a large metric display on the left, trend indicators on the right. Make it look like a high-end venture capital presentation.", "framework": "react", "styling": "tailwind"
}

v0 generates fully functional slide components. These aren't static images; they're responsive, editable components you can customise further in your presentation tool.

Step 5: Assemble into Presentation Format

Create a final step that collates everything: the v0 components, Text2Infographic assets, and narrative text generated by Claude. Use an n8n HTTP request to your internal API or a Google Slides API endpoint to build the presentation programmatically. If you're using Google Slides, create a template presentation and populate it:

CopyPOST https://slides.googleapis.com/v1/presentations/PRESENTATION_ID/batchUpdate
Authorization: Bearer YOUR_GOOGLE_API_KEY
Content-Type: application/json { "requests": [ { "createSlide": { "insertIndex": 0, "slideLayout": "layouts/LAYOUT_ID" } }, { "insertText": { "objectId": "SHAPE_ID", "insertionIndex": 0, "text": "Financial Summary\n\nRevenue: $2.65M\nGrowth: 121% YoY\nRunway: 24 months" } } ]
}

The final step outputs a shareable Google Slides link or downloadable PowerPoint file.

Orchestration in n8n

Wire this all together in n8n using these nodes: 1. Webhook node (input) 2. Chat With PDF node (call API) 3. Claude node (market analysis) 4. Text2Infographic node (charts) 5. v0 node (slide generation) 6. Google Slides node (assembly) 7. Send email node (notification to founder) Set error handling between each node. If Chat With PDF fails to extract clean JSON, trigger a fallback that asks the founder to format data differently. If infographics fail to generate, send a notification asking for simpler data.

The Manual Alternative

If you prefer more control over each slide, use MindStudio to build a custom agent instead. MindStudio lets you create a no-code AI agent that guides founders through deck creation step by step. You design the prompts, choose which models to use (Claude Opus 4.6 for analysis, GPT-4o for copywriting), and structure the conversation flow. The advantage: more founder input at each stage, better customisation of narrative. The disadvantage: slower and more hands-on. Use this approach if pitch decks have significant variation or require highly specific brand voice.

Pro Tips

Watch for PDF parsing failures.

Not all financial PDFs have clean formatting.

Establish a validation rule in n8n that checks whether Chat With PDF extracted numbers in the expected format. If extraction confidence is below 85%, flag it for manual review before proceeding. Otherwise, you might build infographics on bad data.

Rate-limit your API calls.

Text2Infographic and v0 can handle five to ten requests per minute without throttling. If you're processing multiple decks simultaneously, add delays between calls or you'll hit their rate limits and the workflow will stall. n8n's delay nodes help here.

Use GPT-4o mini for copywriting refinement.

After Claude Opus 4.6 generates narrative text for market slides, pass it through GPT-4o mini to tighten language and remove jargon. It's faster and cheaper than using Opus for this secondary task. The quality difference is minimal.

Store infographic versions.

Text2Infographic can generate the same data in multiple styles. Generate both a dark version and light version, then store both. Founders often want to experiment with visual style; having both ready saves regeneration time.

Cache financial extractions.

If the same founder uploads slightly modified versions of their financial statement, cache the previous extraction in n8n. Use a simple hash of the PDF to detect duplicates. This saves API calls and speeds up re-runs.

Cost Breakdown