Every sourcing tool shows every fund the same companies. This engine is different: it reads your deal history — every pursue, every pass — works out what your firm actually buys, and gets the right owners to reply. Fully operated. You take the conversations.
Anyone can buy 5,000 home-services companies. The hard part is getting a 60-year-old owner to write back. Watch the engine source, screen, and reach out, then teach it your own thesis and watch it adjust.
The demo above shows one loop clicking. Under the hood there are three, and they all start from your own deal history — not our database.
Every yes and no in your deal history gets distilled into a written taste document — hard criteria, soft preferences, anti-patterns — each one evidenced by your own past deals. You read it. You correct it. It's yours.
Every company you've ever judged lives in the engine's memory. Each new candidate is compared to its nearest look-alikes: resembles three you pursued, one you passed on. Your history does the screening.
An analyst model scores every candidate against your taste document and writes down its reasons. The reasons become the outreach — specific to the owner, impossible to mistake for a mail merge. That's why they reply.
Every verdict you give, every reply that lands, every call that books flows back into all three. The engine is at its dumbest today.
You don't have a sourcing problem. You have a reply problem. Here's the part that's invisible until you've done it: the same outreach gets wildly different results depending on whether it lands.
The demo above uses invented companies. The capability behind it doesn't.
We build and run the whole origination engine — sourcing, screening, deliverability, outreach — tuned to your thesis and trained by your own decisions. The demo above runs one vertical; the engine runs whatever you buy. You take the owner conversations.