Skip to main content
Governance is where raw column names become semantic. Xura auto-classifies on upload; you review and refine.

​
Column classification

For each column, Xura infers:
  • Meaning — Email Address, Deal Stage, Currency Amount, Person Name, ISO Date, etc.
  • Type — text, number, currency, date, boolean, categorical.
  • Status — Confirmed (green), Needs input (amber), Auto-suggested (blue).

​
When to intervene

  • Ambiguous names — cd in a pharma dataset could be “CD-ROM”, “Change Directory”, or “Circular Dichroism”. Only you know.
  • Category vs free text — Xura might mark a column as free text when it’s actually a fixed set of stages.
  • Custom units — non-standard currency, custom scales, etc.
Click a row → Inspector opens → edit meaning + type + notes → Confirm.

​
”Accept all AI suggestions”

Banner at the top when N columns need input. Bulk-accepts Xura’s best guesses. Fast for well-structured data; risky for anything unusual.

​
Why this matters

Bad governance = bad queries. If Xura thinks a stage column is free text, it won’t group by stage properly. If it doesn’t know a column is currency, it won’t format sums correctly. Every minute you spend here pays off in dashboard + flow accuracy.