Mickey Studio · AI workflows and analytics
One engineer, a team's worth of tickets
- Client
- International electric-vehicle manufacturer
- Project
- Ticket pipeline for a small marketing web team
- Role
- Sole engineer
- of shipped tickets came through AI agents in the first eight weeks
- ~57%
- of merges approved by a human
- 100%
The problem
One engineer, a long backlog and a constant stream of small fixes.
What I did
I built a ticket pipeline. Requests become scoped tickets, an AI agent works each one in isolation and opens a pull request, then reviews its own work. A human approves every merge.
A weekly retro looks at what the reviews caught and updates the agent's rules.
The result
Over half of shipped work came through the pipeline within two months, with no unreviewed code going live. The rules got better after every retro.
Let's work together
Got a similar problem? Send me a few lines about what you're working on and we'll take it from there.
hi@mickey.studio