Consecutive weekly digests
Every Tuesday at 9 AM ET since 26 May. No missed week.
Evaluated, down to top matches
Scored against a four-tier acquisition mandate.
Hours spent searching, per week
The scan, the scoring and the email all run unattended.
What changed
Deal flow stopped being something you go and find
Every week the system pulls listings from three marketplaces across twelve states and watches a fourth channel for things that never reach a marketplace at all. Structural filters drop anything outside the mandate: wrong price band, wrong geography, already under contract. Only what survives gets read by the model.
Over 22 weeks, 218 listings cleared those filters and were scored against a four-tier acquisition mandate. 216 of them were written into digests as cards, each with the business, the location, the asking price where it was disclosed and a reason for its score. 71 fell inside the target industries.
The owner’s job went from searching to reading one email a week, and deciding which one or two were worth a call.
Where the good ones came from
The best deals were not on the marketplaces
The marketplaces supply the volume. They are also where every other buyer is looking, which is why the system watches a second layer: news, alerts and public posts, the places a seller turns up before a broker does.
That layer produced 10 of the 13 top matches. Among them: two owners posting directly in Facebook groups that they were selling, a soil blending and plant nursery advertised on Craigslist with no broker involved, and a listing that surfaced in alerts before it appeared in any marketplace feed. None of those would have been seen by a marketplace search.
This is the part a broker relationship cannot replicate, and the reason the off-market layer earns its place in the stack.
Why it is trusted
A scorer that mostly says no
Of the 216 cards delivered, 131 scored a 1 or a 2. Thirteen scored 7 or above. Almost nothing lands in the middle, because the model is asked to take a position rather than hedge.
That ratio is the product working. A system that rates everything a seven is just a second inbox to ignore. The low scores still appear, below the fold, so the buyer can see what was rejected and why, which is what makes the top of the email worth reading.
All 13 top matches were in Florida, the priority state, and inside the target price band. 147 of the 216 cards arrived with a phone number attached, so a decision could become a call the same morning.
Their AI opportunity mapping engine uncovered business opportunities we couldn’t see before.
How these numbers were counted
Every figure on this page comes from the 22 production digests sent between 26 May and 20 Oct 2026, parsed and reconciled three ways. Upstream scan volume is not quoted as a number here because the exact denominator has not been reconciled against the scraping console. The system runs on free and pay-per-use tiers for a few dollars a month, and every account it depends on, from the database to the model usage, sits under the client’s own organisation rather than ours.