Finding the right person for a warm intro usually means hours on LinkedIn. You scroll profiles, guess who knows who, and send awkward “hey, do you know anyone at…” messages that mostly go nowhere. I’ve done that more times than I’d like to admit. So when I saw an AI professional call a tool the “ChatGPT to find people,” I read the whole post twice.
The tool is called Happenstance. The author shared a step-by-step workflow that connects it to Claude. With one search, they covered a network of 807,000 people. The original poster says that’s something they couldn’t do with LinkedIn alone.
🔍 What Happenstance Actually Does
Happenstance is an AI-powered people search engine built on top of your own relationships. Standard LinkedIn search makes you fiddle with keyword filters. Here you describe the person you want in plain English, and the AI digs through your connections and your friends’ connections to find them.
The clever part is the Claude connector. Once Happenstance is plugged into Claude, you can run searches, read the results, and write intro requests in one conversation. Searching and outreach stop being two separate jobs.
🛠️ The 10-Step Workflow
Here’s the exact process the creator walked through, with a quick note on why each step matters.
- Create a free account. It’s free and takes about a minute, so you can test it with nothing at stake.
- Upload your LinkedIn connections. This is your base data. Happenstance needs to know who you know before it can search.
- Connect Gmail, Instagram and X. Plenty of your real relationships live outside LinkedIn, so this widens the net a lot.
- Add friends so you can search their networks too. This is where the reach grows fast. Your friends’ contacts become people you can reach through a warm intro.
- Open Claude and go to Settings > Connectors. Connectors let Claude talk to outside tools and pull in live data.
- Add the Happenstance connector to Claude. Now Claude can search your whole extended network from inside a chat.
- Paste the search prompt into Claude. The exact wording is in the next section. It names a friend’s network, a location, a type of role, and a specific need.
- Open the list Claude gives you. You’ll see each person’s role, your mutual connections, and even their recent posts. The author called that last part “crazy,” and I agree. Recent posts give you instant context for a personal message.
- Paste the follow-up prompt. Claude turns the list into a friendly intro request with names and LinkedIn links.
- Copy and send it to your friend on LinkedIn or WhatsApp. Your friend gets a clear, easy ask, so the intro is more likely to happen.
💬 Prompts of the Day
These are the two prompts the original poster used, exactly as written. Swap in your own details where you see brackets.
The search prompt:
“Find people in [your friend’s name]’s network, in New York or San Francisco, who run portfolios of companies (PE, VC, manufacturing) and need AI training.”
The intro request prompt:
“Write a friendly message to Noam asking for intros, with names and LinkedIn links.”
Look at how specific the first prompt is. It sets a location, an industry type (PE, VC, manufacturing) and a clear need (AI training). That level of detail is what turns a huge network into a short list you can actually use.
📊 The Results Are Wild
The numbers the LinkedIn user shared are what really got me. One search covered:
- 807,000 people in total across the network
- 170,000 direct connections
- 370,000 more reached through friends’ networks
The results included CTOs, founders and holding companies. And every person on the list had a mutual friend ready to make an introduction. That’s the difference between a cold DM and a warm handoff.
You can also export everything to CSV or share the list with your team. If you work in sales or partnerships, that means you can drop the results into a spreadsheet or CRM and start tracking outreach right away.
LinkedIn was supposed to work like this.
That one line from the post says a lot. LinkedIn promised that your network would open doors. In practice, finding second-degree connections who match a specific profile is clunky and slow. This workflow gets much closer to that original promise.
💡 Tips to Get the Most Out of It
Now that you’ve seen the process, here are a few ways I’d use it so the outreach doesn’t fall flat:
- Be specific in your search. Include location, role, industry and the problem they likely have. Vague prompts give you vague lists.
- Ask before you search a friend’s network. A quick heads-up keeps the relationship healthy and makes them happier to help later.
- Use their recent posts. Mention something the person shared lately. It shows you did your homework.
- Keep intro requests short. Your friend is doing you a favor. Give them a message they can forward in seconds.
- Don’t over-ask. Pick your best three to five matches per friend instead of sending a list of 40 names.
🎯 Who Can Use This
The author’s example targets people who need AI training, but the same workflow fits plenty of goals:
- Sales teams looking for decision-makers with a warm path in
- Founders raising money and hunting for investors through mutual friends
- Recruiters and hiring managers searching for specific talent in trusted circles
- Consultants and freelancers looking for their next client
- Job seekers who want a referral instead of applying cold
I think this is a good example of where AI tools are heading. They’re less about making content and more about connecting the dots between things you already have, like your contacts, your friends and your goals. Plugging tools like Happenstance into Claude through connectors is turning chat assistants into real working tools.
If you’ve been treating your network like a dusty contact list, this is a great reason to put it to work. Check out the full LinkedIn post for the original walkthrough and more details from the creator.