Documentation

Your first change

Apply one small change end to end, so the confirmation, snapshot and preview are familiar before you need them.

This walkthrough makes a deliberately small change across a single process. The point is not the change — it is seeing what Charter does around it, before you run something larger.

Do this in a test collection if you have one. Charter writes real configuration to real projects, and Azure DevOps offers no undo of its own.

1. Open the hub

Organisation settings → Charter → Processes.

Pick one inherited process and one work item type within it. Charter reads the live definition and shows its fields, states, rules and layout.

2. Make a visible change

Something you can confirm by eye afterwards and reverse easily — for example, move a field into a different group on the form, or set a field's default value.

3. Save, and read the confirmation

Saving does not write anything yet. Charter first shows you:

The blast radius — how many processes, work item types, projects and checks the change reaches. This is a count, and it is the number people underestimate.
A summary of what will change, per process. Processes that have drifted apart will show different entries, because the same configuration lands differently depending on where each one started.

Read the process names. Applying to "all processes" is a common and reasonable thing to do, and it is also how one careless save reaches further than intended.

4. Confirm

Two files download automatically before anything is written:

charter-snapshot-before-<timestamp>.json — the live configuration of exactly the resources about to be touched. Keep it. Re-importing it is how you undo.
charter-apply-preview-<date>.md — the full written account of the change, process by process and project by project. It is more detail than the dialog can hold, and it is the record you can attach to a change ticket.

5. Check the result

Charter reloads the live state after a save, so the hub shows what the server now holds rather than what you asked for. If a write failed for one resource, it is reported individually and the others still went through.