A few days ago I shared how GitHub Copilot shipped a complete feature while I was shopping groceries. That story got a lot of attention and one question kept coming up:
“How exactly did Copilot know what to build?”
The truth is, it all starts with how you write your tickets.
From vague ideas to precise prompts
I’ve been experimenting with a small GPT that acts as a Product Issue Documentation Assistant. You give it a rough feature idea, a bug report, or a product improvement. In return, it gives you a perfectly formatted GitHub issue, ready for Copilot or any engineer to pick up.
Each ticket includes:
- A clear summary, problem, goal, and acceptance criteria
- Consistent markdown structure (no fluff)
- Optional UX copy blocks in both English and German
- Enough context for Copilot to start coding instantly
Think of it as a bridge between product thinking and AI-assisted development.
Why it matters
Writing a great ticket is half the work of building a great feature. And when AI tools like Copilot can now interpret those tickets directly, the quality of your documentation becomes the new developer interface.
This GPT helps standardize that step, so your team (or just you) can move faster, with more clarity and fewer handoffs.
Try it yourself
You can try the GPT here: 👉 Issue Ticket Writer GPT
If Copilot can code while you’re shopping, this GPT makes sure it knows exactly what to build.
