Businesses on their way down, and how much of the work AI could take over
A declining business is cheap for a reason. The question worth asking is whether the reason is a structural death sentence — a platform or an algorithm walked away — or just work that costs too much to do by hand. Only the second kind is worth buying.
Every listing here is scored on five signals, each one traceable to a number on the listing page. The heaviest signal is the last one: whether anyone in our index of — cases has actually made money doing this with AI. No precedent, no score — the card says "unproven" and means it.
Five signals, 100 points. Each one comes from a number on the listing, and each card shows you exactly which signals fired and why. It is not a valuation and not investment advice. It answers one narrow question: how much of this business is work that AI can do now, and has anyone actually done it.
Two things override the points entirely. No precedent in the index → "unproven", however high the score: without a case we have no evidence, and claiming potential would be making it up. Profit and traffic both halved → "looks like a corpse": that is a structural death, and a retrofit does not bring it back.
Listings marked Empire Flippers are read from their public marketplace
page, which they publish for indexing; we never touch the individual listing pages, which they mark
noindex. Everything else is entered by hand from a listing someone was reading anyway —
those numbers are transcribed, not verified. Every card links back to the source.