Create a Cleansing Pipeline

Steps

1. Go to Data Management → Pipelines → New → Cleansing Pipeline.
2. Name your pipeline and set Visibility:
   • Private (only you/admins) or Public (all workspace users).
3. Add data source: upload CSV/XLSX (single or multi-sheet).
   • No file size limit.
   • For multi-sheet files, ensure column headers are identical across sheets.
   • No prep needed—drop raw procurement data as-is; Sensecloud auto-identifies spend, suppliers, contracts, etc.
4. Auto-detect & map: the wizard suggests column mappings and dataset types. Confirm or adjust if needed (you can set unique keys, default currency, or date hints).
5. Save & Run: the pipeline executes and auto classification is triggered (supplier normalization, category mapping, dedupe, and basic validations).
6. Review results: check the run log and QC summary; optionally resolve any ambiguous matches in the task queue (all changes are audited).
7. Publish outputs: clean tables are available to agents and for CSV/XLSX export; reuse the saved mapping for future refreshes or schedule the pipeline to run periodically.

Tips

• If your team isn’t ready to collaborate, keep the pipeline Private and switch to Public later.
• After the first successful run, Sensecloud auto-connects downstream agents (e.g., Spend Analysis, Supplier Risk & News) to start generating insights.