360 feedback
This article explains how the 360 feedback feature works in the Moral Fabric app. For the process overview and coaching guide, see the pattern pages: Quarterly Reflections: 360° Feedback and For Coaches: facilitating a Reflection Session .
Your coach
Every round has a coach: a colleague paired with you ahead of time who reads your feedback alongside you and helps make sense of it.
Setup. An owner, admin or manager pairs people in Settings → 360° Feedback . Each person has one default coach, auto-attached when you start a round. You can swap the coach on any round; that also updates your default. More on pairing and reassigning: Coaching pairs .
The coach can: read every response, read the AI summary with you, keep private notes and reminders in the Coach Tools panel, visible only to them, and click "Prepare session" to pre-fill a coaching agenda.
The coach cannot: submit their own evaluation, or read responses before they arrive.
The AI summary
Once responses come in, open the round and click Generate Summary on the AI Summary card. It lists common themes, strengths, areas for growth and questions to take to your coach. You, your coach and Moral Fabric staff can generate it. Who can read it is under Who sees what below.
Mistral AI, a French company that processes data in the EU, writes the summary. The line under the card title says who wrote it and when, such as Generated by Mistral on 24 October 2026 from 4 responses. Older summaries say Generated by Claude.
- Regenerate writes a new summary with Mistral, for example after more responses arrive.
- Try again with Claude rewrites it with Claude, made by Anthropic, when the Mistral version misses the mark. This sends the feedback to Anthropic, which processes data in the US. The button shows only when the current summary came from Mistral.
Optional: quality ratings
Below the open questions, the form has a short list of qualities (e.g. Ownership, Prioritization, Collaboration), each rated on a five-point frequency scale from Rarely to Almost always, plus Not sure. The recipient sees the aggregate as a radar chart and a compact table.
Quality ratings complement the prose: a fast contribution path for colleagues who freeze at a blank text box, and a way to track patterns across rounds. Workspaces can shape their own list or turn the section off. Full mechanics in Qualities in 360 feedback . To see the scores and written feedback across the whole organisation and per team, open 360 Insights .
Manage a round
Everything below sits in the round's ⋯ menu.
- Add Feedbacker to invite someone you forgot.
- Remove one person's invitation if they no longer need to answer. The round shows who was Removed and who Skipped .
- Mark Round Complete to close it before everyone has answered.
- Cancel Request to stop the round. Feedback that already came in stays readable, and Reopen Round brings the round back.
Who sees what
You (recipient)
All responses, AI summary, your reflections
Your coach
Same as you, plus private coach notes
A shared viewer you added
All responses + AI summary; no coach tools
Anyone else
Nothing
Workspace admins can see metadata (who asked whom, when) for safeguarding. They can't read response content. Full access model: Who can see your data .
When it pays off
- Before a 1:1 or a season change. It pulls a real conversation onto the table.
- After a stretch project. Specific moments are still fresh; answers carry weight.
- When something feels off and you can't name it. Open questions surface what you didn't expect to hear.
Run a round when you're genuinely unsure. Signal-to-noise is much higher than when you already know the answer.
Querying feedback from AI tools
If your workspace is connected via MCP , you can ask things like "What themes show up across the last three feedback rounds I ran?" or "Cross-reference the qualities people rated me on against the projects I led this year." The AI sees what you'd see.