How to Automate Investor Updates with AI Using MCP

Writing investor updates is one of those founder tasks that everyone knows is important — and almost nobody does consistently. The data is clear: founders who send regular updates raise follow-on funding at significantly higher rates. But sitting down on the first of every month to compile metrics, draft a narrative, and hit send? That's where discipline breaks down.

What if your AI assistant could do most of it for you?

UpdatePilot now supports the Model Context Protocol (MCP), an open standard that lets AI agents like Claude, Cursor, and other MCP-compatible tools interact directly with your investor update workflow. Instead of logging into a dashboard and filling out forms, you can tell your AI agent: "Create this month's investor update with our latest numbers and send it."

This guide walks through exactly how it works.

What Is MCP?

The Model Context Protocol is an open standard developed by Anthropic that gives AI models a structured way to interact with external tools and services. Think of it as a universal plugin system for AI — instead of each tool building its own custom integration, MCP provides a common interface that any AI agent can use.

When you connect an MCP server to your AI agent, the agent gains access to tools — specific actions it can take, like "add a metric to an update" or "send the update to investors." The agent sees descriptions of each tool and its parameters, then decides which tools to call based on your instructions.

UpdatePilot's MCP server exposes your entire investor update workflow as a set of tools. Your AI agent can create updates, add KPIs, write highlights, generate AI-drafted narratives, and send the final update — all through natural conversation.

Setting Up the UpdatePilot MCP Server

1. Generate an API Key

In UpdatePilot, go to Bot Setup in your company settings. Click Create Key, give it a name (e.g., "Claude Desktop"), and copy the key immediately — it's only shown once. The key starts with upk_ and authenticates your AI agent with your account.

2. Add the MCP Configuration

Add the UpdatePilot MCP server to your AI agent's configuration. For Claude Desktop, add this to your claude_desktop_config.json:

{
  "mcpServers": {
    "updatepilot": {
      "command": "npx",
      "args": ["@updatepilot/mcp-server"],
      "env": {
        "UPDATEPILOT_API_URL": "https://updatepilot.ai",
        "UPDATEPILOT_API_KEY": "upk_your_api_key_here",
        "UPDATEPILOT_COMPANY_ID": "your_company_uuid"
      }
    }
  }
}

Replace the API key and company ID with your actual values (both are available on the Bot Setup page). Restart your AI agent, and the UpdatePilot tools will appear.

3. Verify the Connection

Ask your AI agent: "Use the get_company tool to show me my UpdatePilot company info." If it returns your company name and settings, you're connected.

The Automated Update Workflow

Once connected, here's the end-to-end flow for creating and sending an investor update through your AI agent. This mirrors the same step-by-step process you'd follow in the UpdatePilot dashboard, but driven entirely by conversation.

Step 1: Create the Update

Tell your agent to create an update for the current period:

"Create an investor update for March 2026."

Behind the scenes, the agent calls create_update with the period "2026-03". This initializes the update or returns the existing one if you've already started it.

Step 2: Add Your Metrics

This is where automation shines. Instead of manually entering each KPI, you can paste or describe your numbers:

"Add these metrics: MRR is $47,200 (up from $41,000 last month), burn rate is $62,000, runway is 14 months, and we have 312 customers (was 287)."

The agent translates this into multiple add_metric calls, each with the name, value, unit, previous value, and display name. It handles the formatting and structure so you don't have to think about it.

Step 3: Add Highlights and Lowlights

Share what went well and what didn't:

"Highlights: We closed our first enterprise deal with Meridian Health ($18K ACV), shipped the v2 dashboard that customers have been requesting, and hired a senior backend engineer. Lowlights: Churn ticked up to 4.2% from some legacy customers on the old plan, and our iOS app launch is delayed to next quarter due to App Store review issues."

The agent calls add_highlight and add_lowlight for each item, categorizing them automatically. For lowlights, it captures mitigations when you provide them.

Step 4: Add Asks

Investor updates should include specific ways investors can help:

"Asks: We're looking for a warm intro to the VP of Digital Health at Kaiser Permanente (high urgency), and we'd love advice on structuring our Series A deck."

The agent calls add_ask with the request text, type (intro vs. advice), and urgency level.

Step 5: Generate the Narrative

Here's where UpdatePilot's AI drafting meets your AI agent:

"Generate a draft narrative from all the components."

The agent calls generate_draft, which sends all your metrics, highlights, lowlights, and asks to UpdatePilot's backend AI. It returns a polished, founder-voiced narrative that ties everything together. You can review it, then ask your agent to refine specific sections using edit_narrative.

Step 6: Review and Send

Before sending, review the complete update:

"Show me the full update for March 2026."

The agent calls get_update and displays everything — metrics with month-over-month deltas, the narrative, highlights, lowlights, asks, and your contact list. When you're satisfied:

"Send it."

The agent calls send_update, which delivers the update to every investor on your contact list via email. This action is final — once sent, the update is locked and your streak counter increments.

What You Can Do with the MCP Tools

UpdatePilot's MCP server exposes 12 tools that cover the complete update lifecycle:

You can also use remove_component to delete any metric, highlight, lowlight, ask, or evidence item, and edit_narrative to directly modify the narrative text.

Why Automate Investor Updates?

The biggest barrier to consistent investor communication isn't that founders don't care — it's that the task sits in an awkward zone: important but not urgent, recurring but not habitual. Automation changes the calculus:

• Consistency beats perfection. A sent update is infinitely more valuable than a perfect draft sitting in your todo list. When sending takes 2 minutes of conversation instead of 30 minutes of form-filling, you actually do it every month.
• Your AI agent already has context. If you use an AI assistant throughout the month, it already knows about your wins, challenges, and metrics. The MCP connection lets it channel that context directly into a structured update.
• Investors notice the cadence. Regular updates correlate with higher follow-on funding rates. UpdatePilot tracks your streak — and your investors see it.
• The narrative gets better over time. UpdatePilot learns your company's tone and voice. Each update you send refines the AI's understanding of how you communicate, so drafts improve month over month.

Advanced: Pulling in Data Automatically

Because MCP is an open protocol, you can chain multiple MCP servers together. Imagine this workflow:

1. Your AI agent pulls revenue data from your Stripe MCP server
2. It pulls usage metrics from your analytics MCP server
3. It creates an UpdatePilot update and populates the metrics automatically
4. It drafts the narrative based on the data trends it observed
5. You review and hit send

The entire monthly update process collapses to a single conversation. The data flows directly from your systems into a polished investor email — no spreadsheets, no copy-pasting, no context switching.

Getting Started

If you're already an UpdatePilot user:

1. Go to Company Settings → Bot Setup
2. Create an API key
3. Copy the MCP configuration JSON
4. Add it to your AI agent's config and restart
5. Try: "Create an investor update for this month"

If you're new to UpdatePilot, sign up for free and set up the MCP connection in a few minutes. The free tier includes one update per month — enough to see the automation in action.

Your investors are waiting to hear from you. Now there's no excuse not to send that update.