THE THING WE BUILT:

The AI Ticket Whisperer

Last week, as I putting my kiddo to bed, they asked me, “Papa, what do you like most about your job?”

I brushed their hair away from their face and softly whispered, “writing tickets.”

They did not appreciate the sarcasm.

Traditional ticketing is like playing telephone with your friends, but without the laughter or snacks. You describe a problem. Someone interprets the description. Assigns it to someone else. Adds some labels. Kicks it to the person it was supposed to go to in the first place. It's exhausting.

Which got us thinking: What if instead of trying to make AI fit into our existing processes, we built processes with AI in mind from the start?

So... we built a thing.

How it works:

Picture this: You’ve got your list of tickets in front of you. Instead of playing “20 Questions” with a ticketing system, you simply:

1. Take your ticket description (messy thoughts and all)

2. Share it with our AI assistant

3. Watch it transform into a properly structured ticket

The AI handles the tedious parts:

• Suggesting relevant labels and assignees based on similar past tickets

• Spotting connections to related work

• Converting casual descriptions into clear requirements

• Learning your team's patterns over time

Things we learned from the thing:

1. Cursor (our AI coding assistant) and Linear (our ticketing system) worked together better than we expected. It's like having a really smart autocomplete for your entire workflow, not just your code.

2. The sweet spot is letting AI handle the tedious stuff while keeping humans in control of the important decisions. We tried full automation first — it was too much. We tried minimal AI assistance — it was too little. The Goldilocks middle ground is just right.

3. Development time dropped significantly because tickets arrived pre-organized with clearer requirements. It’s like the tickets are already half-solved when you open them.

4. When you build for AI first, the coordination tax goes down. People spend less time managing tickets and more time solving problems. It’s not just about making existing processes faster — it’s about rethinking how we work together.

P.S. For the technically curious: We built this using Visual Studio Code with the Cursor AI extension, connecting directly to Linear's API. The AI interprets natural language inputs and converts them into structured data using a combination of pattern matching and custom prompts. Want to know more about the technical implementation? Just reply to this email.

Built with ❤️ by Township

You’re receiving this because you signed up for Township’s "We Built a Thing" newsletter. We'll only send these when we, well, build things.