How I Built an AI-Powered Outreach Engine for Newly Funded Startups Using n8n
Finding new consulting or fractional opportunities often comes down to timing. The best moment to reach out to a founder is right after they raise funding and start planning their next phase of hiring and product acceleration.
I wanted a system that could help me identify newly funded startups, research them, and draft outreach that feels personal and relevant. At the same time, I still want to approve everything manually. Authenticity matters. I do not want to fire off automated cold emails.
This post walks through how I built a light agentic flow using n8n, OpenAI, and a personal vector store.
Google Alerts Trigger
I set up a Google Alert for:
seed AND funding
Every morning, Google emails me freshly funded startups.
Then in n8n, I run a cron schedule to pull that email:
0 9 * * 1-5
This checks Gmail on weekdays at 9 AM for the latest alert from:
googlealerts-noreply@google.com
(this is the Google’s automated alert mailer address).
Extracting Article Links
The Google Alert email includes multiple links inside an HTML script block.
In n8n, I added a small Code node to extract the URLs. Conceptually:
const html = $json.html;
const links = extractLinksFromGoogleAlert(html);
return links.map(url => ({ json: { url } }));
This produces a clean list of article URLs.
Extracting Company Funding Details
For each article:
-
Fetch the page
-
Clean the text
-
Send it to OpenAI
-
Receive structured output back
Schema used:
{
"company": "string",
"industry": "string",
"funding_round": "string",
"funding_amount": "string",
"investors": ["string"],
"founders": [
{ "name": "string", "title": "string" }
],
"summary": "string",
"source_url": "string"
}
This converts press releases into structured business context.
Personal Knowledge Layer: Vector Store
To make messages more personal, I uploaded:
-
My resume
-
Project writeups
-
Notes on startup & AI builds
-
Cloud architecture work
into an OpenAI vector store.
When processing a startup, the flow queries my vector store:
Find relevant experience based on this industry and summary.
This provides relevant snippets from my background to ground the outreach message.
Generating the Outreach Email
Then n8n sends a final prompt to OpenAI:
You are an expert startup operator and technical leader. Write a short, respectful founder outreach email based on this funding event.
Requirements:
- Congratulate the founder
- Reference the funding, investors, and mission
- Summarize why their vision is compelling
- Tie in my background using the experience summary provided
- Keep the tone friendly, concise, and professional
- No aggressive sales language
- Return only JSON with keys "subject" and "body"
Here is the startup info:
Here is my relevant experience from vector search:
Respond in this JSON format:
{
"subject": "...",
"body": "..."
}
Format expected:
{
"subject": "string",
"body": "string"
}
Example mapping in n8n:
Subject:
Body:
Saving as a Gmail Draft
Instead of auto-sending, I create a Gmail draft. I still review and send manually.
The workflow does the research and writing. I stay responsible for final touches and delivery.
This keeps the outreach thoughtful, not spammy.
This Was a Learning Project
Important note:
This is a demo project and a learning exercise. I am not claiming to be a cold-outreach expert.
My goals here were to:
-
Explore multi-step automation
-
Add personal context via vector search
-
Learn how to chain agents in n8n
-
Keep a human in the loop
It helped me understand how agentic workflows support solo builders.
Final Thoughts
This workflow gives me a clean, sustainable way to surface interesting startups and prepare relevant outreach without spending hours researching manually.
It is a small but meaningful example of how AI automation can help indie builders and fractional engineers without replacing personal connection.
If you’d like to see the n8n workflow, feel free to reach out. You can always find me on Twitter at @karancito.
