An intern at Listen set an AI agent loose on the full founder playbook with no starting product in mind, letting it run 2,000 interviews, test roughly 100 concepts, build and refine a live app, then launch paid ads that brought in hundreds of real customers for the resulting stylist tool.
super impressive example of combining machines-asking-humans in order to efficiently direct use of compute against a goal
Our intern just built the first zero-person company.
Listen's agent ran a loop: - Interview users - Build - Test with real people - Fix issues - Repeat
2,000 interviews and 100 concepts later: an app with 100s of paying customers.
Here’s how it works:
The agent handled interviews, iteration, and even ad testing on its own, yet needed people to open the Meta account, connect Stripe, and approve spend, so the zero-person claim stops short of total independence.
The final app collected over a thousand dollars from hundreds of customers, yet ad costs ran higher and the experience stayed rough enough that profitability stayed out of reach in this first run.
Many users are excited by an AI agent building a zero-person company with hundreds of paying customers as a glimpse of the future, while others dismiss the product as low-quality AI slop or argue it still requires human oversight.
No Digg Deeper questions have been answered for this story yet.
the future of work is interning for an LLM
Our intern just built the first zero-person company.
Listen's agent ran a loop: - Interview users - Build - Test with real people - Fix issues - Repeat
2,000 interviews and 100 concepts later: an app with 100s of paying customers.
Here’s how it works:
the next year will all be self-improving for [insert thing]
Our intern just built the first zero-person company.
Listen's agent ran a loop: - Interview users - Build - Test with real people - Fix issues - Repeat
2,000 interviews and 100 concepts later: an app with 100s of paying customers.
Here’s how it works:
"our intern built the first zero-person company"
wait does the intern not count as a person
Our intern Veer was so proud when the agent got its first customer
Our intern just built the first zero-person company.
Listen's agent ran a loop: - Interview users - Build - Test with real people - Fix issues - Repeat
2,000 interviews and 100 concepts later: an app with 100s of paying customers.
Here’s how it works:
Our intern’s prompt: “You are an autonomous founder-agent and your goal is to create a viral app, using Listen Labs interviews”.
The agent then:
1. Ran discovery studies, used subagents to explore 100 concepts in parallel, and narrowed it to one.
tbh, the product is really janky. The image model distorts your face and the UI is clearly designed by AI.
But we’re excited to see *the start of* what’s possible when you combine compute with a constant feedback loop with real users.
We’ll continue stress testing the boundaries of research and report back on what we learn.
Read our methodology here: http://listenlabs.ai/blog/closed-loop
5. Tested messaging and launched ads (via Meta MCP). It got hundreds of paying users.
You can try the app at: http://stylefits.ai.
@itsalfredw I'm glad it did not end up with a gambling site haha
2. Built the first version (LooksMax) that scored people’s looks. NPS was -38. Everyone thought it was shallow and not actionable.
4. Looped usability, trust, and pricing studies and added improvements: privacy guarantee, a free report, direct shopping links. Then it deployed the live site with Stripe payments.
3. Interviewed more users and rebuilt the product as StyleFits. Users upload a photo and get outfit, haircut, and color palette recommendations.
@itsalfredw Amazing!!
@itsalfredw teams cooking! next goal: zero person self governing, cash flow positive entity. EntrepreneurBench
@itsalfredw give this man a raise
@florian_jue @itsalfredw I believe the term of art is "prediction market."
@itsalfredw What's its NPS
@itsalfredw hey man can you stop tweeting because every time i see you guys cook i feel retarded
@itsalfredw When can we get the Listen MCP to spin up interviews??
@deedydas @andonlabs is building this!
@itsalfredw future of software
