Why I built Shipped or Stalled
I spend most of my time with leaders at large companies who are trying to get AI out of the pilot stage. Sooner or later, almost all of them ask me the same thing: is anyone actually getting agents into production, or is it all demos?
Until now there was no single place with a straight, sourced answer. There is no shortage of announcements. A bank launches an assistant, a retailer unveils an agent, a vendor publishes a case study. Then the story stops. Very little coverage goes back to check what happened next.
So I built Shipped or Stalled. It is a free, public registry of named AI agent projects at big companies. Every project gets one of four tags: Pilot, Production, Scaled, or Killed. Every tag links to a public source. If the source does not make the status clear, the project stays out.
Today it tracks 144 projects: 76 in production, 29 scaled, 20 killed, and 19 still in pilot.
The killed projects matter
The killed projects are as useful as the wins, because they show where things break.
This is the argument I make on this site, turned into data. Most companies have an AI strategy. Few have production. The gap between a pilot and a production-ready system is where organizations go to die, and I wanted a scoreboard that shows that gap honestly.
AI theatre gets a press release. AI-Native work gets a status that holds up.
The rules
- Every status links to a public source, and we use only what that source says.
- Status is never for sale. Sponsors are labeled and never affect which projects appear or what status they get.
- No consulting leads and no partner listings.
- Shipped or Stalled is made at TribalScale.
Help me keep it honest
Know a project that is missing, or one whose status has changed? Send the public source through the submit page. Every submission is checked against its source before it appears. A weekly newsletter is coming soon.
If you are deciding which AI agent projects to fund, scale, or stop, I built this for you. Take a look, and tell me what we got wrong.