Use AI import safely with validation gates that protect data quality before production workflows.
- Structured source data prepared
- Human reviewer assigned
- Use consistent column conventions and avoid ambiguous abbreviations.
- Split very large imports into manageable batches.
- Upload/paste data in import flows.
- Review parsed entities before committing.
- Check duplicate dancers/routines/studios and merge manually when needed.
- Run post-import spot checks in downstream tabs (dancers, routines, venues).
- No critical duplicates created by batch import.
- Imported category values match competition taxonomy.
- Downstream tabs render imported data without schema gaps.
AI parsing creates malformed rows
- Clean source format and retry smaller batch with explicit headers.
Imported data cannot be scheduled/scored
- Backfill required metadata (style, level, age, performers, venue where relevant).