Creating an investor pitch deck requires more than placing text into slides. Teams need to understand the company’s business model, research the market, assess competitors, shape the problem and solution narrative, prepare financial visuals, and present everything in a clean format investors can review.
This use case explores how a connected AI automation can take company information from a spreadsheet, conduct structured market and competitive research, generate an investor narrative, populate a Google Slides template, create a branded financial chart, and log completion metrics for the team.
The Problem
Putting together a genuinely investor-ready pitch deck can take days of work per company. It often requires researched market sizing, a credible competitive landscape, SWOT analysis, a persuasive problem, solution, and opportunity narrative, and clean financial charts.
For a venture programme, accelerator, or startup support team working with multiple companies, this process becomes difficult to manage manually every week. Without automation, teams may spend significant time researching, writing, formatting, editing slides, and updating financial visuals for each startup.
The Solution
The AI Investor Pitch Deck Generator is an end-to-end pitch deck automation workflow.
It pulls a pending company’s business-model data from a spreadsheet, runs market and competitive research through specialised AI agents, writes an investor narrative, populates a Google Slides template, generates a branded financial chart, places it into the deck, and logs time-saved and completion metrics.
The workflow also includes an on-demand reporting endpoint that summarises deck-generation performance for the team.
How the Agent Works
- Receive company information. The workflow reads a company marked as pending and pulls its business model, value proposition, customer segments, financials, and key metrics.
- Research the market and competitors. AI agents gather market-size, growth, and competitive information from credible sources.
- Create the investor narrative. The workflow generates a structured problem, solution, and opportunity narrative using the company’s own data and the gathered research.
- Populate the pitch deck template. A Google Slides template is copied and filled with the company’s information, research, narrative, business model, and competitive analysis.
- Generate and insert a financial chart. Financial projection data is converted into a branded chart and placed into the correct slide position.
- Log results and report performance. The workflow saves the deck link, records completion metrics, and supports on-demand KPI reporting for the team.
Technical Workflow
1. Weekly intake
A schedule trigger reads the next company marked “Pending” in a “Business Model Information” sheet, used for Divverse’s Venture & Talent Development programme.
It pulls the full business-model canvas, including value proposition, revenue streams, key partners, key resources, key activities, customer segments, channels, cost structure, competitive advantage, financials, and key metrics.
2. Session setup
A code node generates a unique session ID for the run, used to key conversational memory across the research agents.
3. Market Research Orchestrator
A Gemini-powered orchestrator agent is explicitly forbidden from researching anything itself. Its only job is to hand structured tasks to two sub-agent tools, receive their JSON, and assemble one unified four-key response covering market size and growth, competitive analysis, and matching markdown summaries under a strict schema with inline-citation formatting rules.
4. Market Size and Growth sub-agent
This sub-agent uses a Tavily web-search tool restricted to credible sources, including government, academic, major research firms, and established financial press.
It also uses an explicit exclusion list, including Wikipedia, forums, blogs, content farms, and social media. The agent generates current and forecasted market size, CAGR, growth drivers, challenges, and regional trends, all time-windowed to the past two years.
5. Competitive Analysis sub-agent
This sub-agent uses the same Tavily tool and source-credibility rules to identify 4–5 named competitors and build a full business-model-canvas-plus-SWOT profile for each.
The profile includes overview, revenue or funding, market share, business model, value proposition, key partners, key resources, key activities, customer segments, channels, and strengths, weaknesses, opportunities, and threats, with every claim sourced.
6. Parse and flatten research
A code node parses the orchestrator’s combined JSON and flattens it into a single spreadsheet row.
It expands the competitor array into five fixed sets of columns, from name through competitive advantage, so the data can live in a normal sheet.
7. Log research and mark done
The flattened row is written back to the Business Model Information sheet against the same company row, and its status flips from Pending to Done.
8. Generate the investor narrative
A second LLM chain, powered by Gemini and prompted as an “expert startup pitch writer,” is instructed to use only the company’s own data and the freshly gathered market research, never training knowledge.
It produces a strict three-part Problem / Solution / Opportunity narrative, with each section written in 3–5 sentences and returned in JSON.
9. Parse and log the narrative
A code node strips markdown links and code fences, parses the JSON, and falls back to regex extraction if parsing fails.
The three narrative sections are then appended back to the company’s sheet row.
10. Clone the deck template
A Google Drive copy operation duplicates a master “Investor Pitch” Slides template, naming the new copy after the company and today’s date.
11. Populate the entire deck
A single Google Slides node runs dozens of find-and-replace operations across the copied deck.
It replaces fields for company name, industry, region, mission statement, milestones, the Problem / Solution / Opportunity narrative, how-it-works sections, the full business model canvas, up to four competitors with strengths, weaknesses, opportunities, and threats, and go-to-market details.
This turns a static template into a company-specific deck in one pass.
12. Pull financial projections
A separate “Financial Projections” sheet tab is read for the same company.
The data includes revenue, expenses, cash flow, MRR, COGS, gross profit, and net income by year or quarter, then gets flattened into a semicolon-delimited dataset string.
13. AI-generated financial chart
An agent, GPT-4o-mini with a Think tool and a strict structured-output schema, turns the raw financial dataset into a valid, minimally sized Chart.js configuration.
The chart is a revenue bar chart with an expenses line overlay. It enforces an exact brand colour and font scheme, including deep blue and silver bars, a consistent blue accent across titles, legend, and axis labels, and Lato typeface to match the deck’s geometric investor pitch theme.
14. Render and place the chart
The Chart.js configuration is POSTed to QuickChart to generate a hosted image.
The target slide is fetched from the Slides API to locate a tagged “ChartPlaceholder” element and read its exact size and position. A Slides batchUpdate then inserts the rendered chart image into that exact spot, so the chart visually replaces the placeholder at pixel-accurate scale.
15. Finalise and log the deck link
The finished presentation’s shareable URL is written back to the company’s sheet row.
16. Self-reported agent metrics
A code node computes the automation’s own ROI for the run.
It calculates minutes saved compared with an 8-hour manual benchmark, how many market-research data points were gathered, a slide-completion rate, and a conversation-efficiency score based on how many back-and-forth turns the research agents needed.
All metrics are appended to a dedicated “Agent KPIs” sheet.
17. On-demand KPI reporting
A separate webhook accepts an optional date range, pulls every logged KPI row, fuzzy-matches column names to tolerate naming drift, normalises percentage values that may be stored as either 0–1 or 0–100, aggregates totals and averages across the window, and responds with a single auto-written narrative sentence.
The response summarises decks generated, time saved, data points gathered, and completion and efficiency rates for that period.
Technology and Integrations
Built with: n8n, Google Gemini, OpenAI GPT-4o-mini, Tavily Search, Google Sheets, Google Slides API, Google Drive, QuickChart, webhooks, and a KPI reporting workflow.
Outcome
The AI Investor Pitch Deck Generator creates a structured way to produce investor-ready pitch decks from company data, market research, competitive analysis, financial projections, and a branded slide template.
Instead of manually researching, writing, formatting, and editing every deck, the workflow can move from a company marked Pending to a populated Google Slides presentation with a sourced research base, investor narrative, SWOT-backed competitive section, and branded financial chart.
The workflow also tracks its own performance, including time saved, data points gathered, slide-completion rate, and efficiency metrics, giving the programme team a running view of automation ROI.
Build Custom AI Automation for Your Business
The AI Investor Pitch Deck Generator is one example of how AI automation can support research, content generation, reporting, document creation, and business-development workflows.
We design and build custom AI agents, automation workflows, internal tools, and connected systems for a wide range of business processes.
From sales, marketing, recruitment, customer support, and reporting to finance, operations, data, and internal team workflows, each solution is designed around the way your business works.
Request a Quote for a Custom AI Automation Solution to get started.