← All use cases

Analyze product feedback across tools

Gather themes from Slack conversations, survey exports, and issue backlogs, then list what to do next.

Feed ZeroTwo Work feedback gathered from Slack, surveys, issue trackers, support tickets, or research notes, and it clusters recurring pain points in a Sheet or Doc you can review, backed by evidence, design takeaways, open questions, and next steps.

Easy
30m

Skills & Plugins

  • Slack
    Reads feedback from approved channels or from thread links you give it.
  • GitHub
    Reads issues, pull request comments, and discussions.
  • Linear
    Reads bug and feature queues.
  • Google Drive
    Reads feedback docs and exports, and can create the Sheet or Doc that holds the results.

Starter prompt

Pull the last 60 days of feedback about our onboarding flow from the #product-feedback Slack channel, the Linear issues labeled onboarding, and the attached survey export. Group it into recurring problems, and for each one give a short description, how often it appears per source, two or three linked examples, and your confidence. Put the result in a new Google Sheet, and draft a short summary for the design team that I can review, without posting or assigning anything.

Before you start

Tell Work which sources, which product area and what date range to use; it clusters recurring problems for team review. Attach local exports to the chat, or run the task on the desktop app to read files on your machine.

What to expect

A first pass surfaced three repeated problems, with mention counts across sources, then a follow-up split one theme in two. Counts reflect a small sample, not incidence.

ZeroTwo Work

How it works

Provide sources, request themes with linked evidence, record affected users, confidence and open questions, and review before drafting anything outward.

Turn a reviewed theme into the next draft

Split broad themes, add evidence or draft an update for a named audience, unposted.

Keep a feedback channel current

Create a scheduled task that rereads the sources and drafts a summary. Scheduled runs have nobody to approve actions and run with full access, so a "never post or assign" line in the prompt is only an instruction, not a guardrail. If the task could post or assign, point it at read-only sources or a connector you have limited, and check the first runs. See Permissions.

Related use cases