All selected work
08 / Case study Agent · Client feedback

An agent that turns scattered client feedback into tickets, and acts only where it is allowed to.

Client feedback arrived as scattered comments across the messaging tool, the QA platform and email. An agent rebuilds the threads, files itemized tickets, and acts only on work inside the scope it was given.

Where it stands Built and used · accuracy not measured
System / 08Set up the agent and defined what it may act on
01 Problem → decisions → result
The problem

What needed
solving.

On client projects, feedback came in at high volume from several places at once (the client messaging tool, the QA platform and email) as discrete comments rather than threads.

The calls that mattered

What I decided,
and why.

  1. 01

    Rebuild the threads. The agent reconstructs coherent conversations from discrete comments.

  2. 02

    Consolidate them into itemized tickets in the issue tracker.

  3. 03

    Bounded autonomy. The agent may execute tickets inside its defined scope; anything outside it stays with a person.

Where it landed

What shipped,
and what isn’t proven.

It was built and used. Not proven: how accurately it rebuilt threads, how often its own fixes needed rework, and whether it saved any time. None of that was measured.

Stack
ClaudeJira
Next case study A multi-step agent chain where every image has to land on the right product.