How to find investors for your startup with Claude
Human in the loop
Without the human, none of it works. That is the Saverflow AI investor outreach strategy.
From a list to a reply
The same playbook we run for B2B clients, except the deal is a round and the pipeline is the raise.
investor-icp → investor-sourcing → investor-scoring → investor-messaging → campaign-qa
Most fundraising advice is: get warm intros, polish the deck, pray
We treat investors as an ICP instead. Same playbook we run for B2B clients, except the deal is a round and the pipeline is the raise: they have a mandate, a cheque size, a geography and a reason to take the call. Define them, find them, score them, reach out.
We run this for companies raising now. One is raising €25M and has closed €5M with this system.
A real project, not a made-up company
What happens in the hour
- How much are you raising?
- Who have you already talked to?
- Which type of investor do you want?
- Where are you based, and where is your market?
- Which countries do you want investors from?
- Any region you want to avoid?
- How do you reach investors today, and what has worked?
- Which events are you going to?
Five minutes from the side that receives the message. Bryant reads what we just wrote and says what makes it get opened, answered — or deleted.
Answer it for your own raise
The same questions, on your company. Type as we go — the answers stay in this browser and nothing is sent anywhere, so the filled page is yours to screenshot at the end.
What we refuse to do
A founder asked us to email 100,000 investors. We said no. You don't get 100,000 shots — you get one bad reputation. Four things we won't do, on this project or on yours:
Volume is not a strategy. A hundred thousand names is a hundred thousand people who will never open the second message. We send to the few hundred whose mandate, ticket size and geography actually match the round.
Nobody owes you an intro, a reply or a cheque. The message has to earn the meeting on its own, which is why it carries evidence — traction, a thesis fit, a reason this fund in particular — instead of a request.
A reply rate you can screenshot is worth nothing if the fund never takes the second call. We optimise for the investor who is still on the list at the next round, not for the number in the dashboard this week.
If the numbers, the market and the amount you're raising disagree with each other, no amount of personalisation saves the campaign. We fix the story before a single message goes out.
Paste this into Claude Code
Strategy is answered, so the build can start. This one prompt runs the first three skills: it writes the ICP, sources from both databases and scores the result. Copy it and run the same thing.
What a skill is
A skill is a recipe for Claude Code. Write it once, run it every time.
A text file. It says: do this job, look here, ignore that, give me the result in this shape.
A prompt typed live gives a different answer every time. A skill gives the same answer every time. Same list, same score, same message.
The rules are written down before we start. So when we say no to an investor, everyone can see why.
Two databases, one list
Sourcing doesn't start from a blank search. Two databases feed the same list, and the room sees which one every name came from.
External · wideWho is running it
What you'll take away
- Fundraising as GTM: investors as an ICP, the round as a pipeline
- Why the work runs on skills, not prompts — and what goes inside one
- Where Claude Code fits — sourcing, scoring, enriching, outreach — and where the human has to stay in the seat
- Why the first message chases a meeting, never money
- LinkedIn as an investor outreach channel, without expensive tools
Who it's for
Founders raising now or within a year. GTM and sales people who want to see Claude Code on a real pipeline, not a demo.
The session has taken place.
Saverflow AI
We help companies scale revenue and operations with AI, without hiring. Fundraising is the same machine pointed at a different buyer. Three systems: