Approve your first definitions
Review what Doram found in your data and decide what to keep.
After the first scan, Doram drafts what each table and column means and how tables join. Low-risk changes go live on their own. Medium-risk ones go live too, and 1 in 10 is spot-checked. Anything that could change an approved number waits for its owner.
The review queue
Open Context → Review to see what's waiting on a person. The tables your team asks about most come first, and each item says why it needs you:
| Reason | Meaning |
|---|---|
| Would break something | Accepting it changes something already in use. |
| Conflicts | It disagrees with something else Doram found. |
| Waiting | It depends on a suggestion that isn't accepted yet. |
| Low confidence | There's little evidence for it. |
Reviewing a table
Each table goes through four steps in order: what the table is and what one row means, what each column holds, how it joins to other tables, and the measures and splits answers use (its semantic model).
Open a suggestion to see the proposed change, the evidence for it, how confident Doram is, and what else depends on it. Evidence can be saved queries that use it, a check of the actual values, or constraints in your database.
Then accept, reject, or ask about it. You can edit the text before accepting. If the change would break something, Doram warns you first, and you can still choose Accept anyway.
You don't have to finish the queue on day one. Start with the tables people ask about most.
Business terms
Some meanings aren't in the database at all, like what counts as an "active customer". Your team writes those in Context → Business, and Doram never guesses them. Adding terms in the app is coming soon.