How Aiolum checks every number
Each of the 110 stories is built the same way. The sources are listed at the end of the story, numbered, and each number in the story points to the sources that support it. A source marked "Read in full" was downloaded and the numbers were checked in its text, not in a summary of it. 324 facts across the site carry grade A, the highest grade.
A number gets grade A when a public company disclosed it in a filing, a press release or an earnings report, or when a payment processor verified it. Grade B means two or more independent reports agree, or a journalist confirmed it with the company. Grade C is a founder claim that no one else verified; a screenshot is still a claim. A value marked calc was computed here from graded facts, and the formula is shown next to the result.
Every story then has the same five parts, so that two stories can be compared: the numbers with their grades, the timeline, why it worked, the failed startups that came before (498 of them across the site), and whether one person can copy the pattern now. The About page describes what the site is and who it is for.
Every story, the same five parts
A success story on its own teaches little. So each entry follows one structure, and the same structure will apply to every story that is added later.
- 01The numbers, graded
- 02The timeline
- 03Why it worked
- 04The dead that came before
- 05Can you copy it
Four labels, and nothing without one
Most online success stories are founder claims, and most of them exist to sell a course. Here every number carries one of four labels.
A public company disclosed it in a press release, a filing or an earnings report, or a payment processor verified it.
Two or more independent reports agree, or a journalist confirmed it with the company.
The founder reported it, and no one else verified it. A screenshot is still a claim.
We computed it from graded facts, and the formula is shown next to the result.
What counts as a one-person AI business
AI-native means that the model output is the product, for example an app builder or an image generator. AI-leveraged means that one person runs a normal product and uses AI tools to do the work of a team. A story qualifies when the founder was alone until the product worked. Most successful founders hire a few people after that point, and the story says when.
Why every story is paired with failures
A success alone can be luck. A success next to five earlier failures of the same idea shows what changed. The failures come from the Loot Drop database of 1,749 dead startups. Each failed company links to its entry there.
Know a verified AI success story?
One person, real revenue, public evidence. Send it, and it gets the same five-part treatment.