Investors win by reaching great founders before they start fundraising openly. We built a pipeline that spots those early signals automatically, so the team spends its time on outreach, not research.
The problem
Two signals matter most. Someone engaging with several VC funds’ posts in a short time is often starting to raise. And alumni of high-value companies quietly starting something new are worth knowing early.
Both are easy to understand and almost impossible to track by hand, across dozens of funds and thousands of interactions. By the time the team heard about good founders, they were already deep into their raise.
What we built
- 01
Watching the signals, around the clock
The system monitors public posts from the fund partners it tracks, sees who engages, and follows alumni of high-value companies for signs of new activity. Each stage runs separately, so it handles volume without overloading the source platforms.
- 02
Rules you can read
Instead of a black-box model, a clear, adjustable rule engine spots founders and stealth-mode signals, filters out recruiters and students, and identifies the likely company, with a confidence score.
- 03
One scoring formula everyone trusts
This was the hardest part. Six different scoring methods had built up over time, each with its own idea of a strong lead, and they contradicted each other. We merged them into one auditable formula that rewards fund diversity, recency, alumni value and seniority.
- 04
A person decides who to contact
Leads move through a lightweight CRM, from new to reviewed, contacted and converted, with reviewer notes. No founder is contacted on an automated score alone.
The result
- Four alert types that rank leads by signal strength, without duplicates
- A pipeline that grows by adding workers, not rewriting logic
- One transparent scoring formula instead of six conflicting ones
- Every flag explainable, with no black-box model
- Outreach that stays human-led
Could this work for you?
Something similar could help if:
- Your best opportunities are hidden in public data no one has time to track
- You need to reach people before your competitors do
- You want automation you can read and adjust, not a black box
