Data & analysisEntity resolutiondata
Review possible duplicate customers without auto-merging
Separates exact matches, plausible matches and risky false positives.

Ready to use
Prompt
Given permitted [CUSTOMER RECORDS] and [MATCH RULES], build a privacy-preserving review plan for duplicate accounts. Normalise names and contact fields in a reversible way, score exact versus fuzzy evidence, and output candidate pairs with reasons and conflicts. Never merge, delete or contact accounts automatically. Include a human decision record and rollback plan for any later approved merge. Check that shared addresses, family members and recycled phone numbers do not cause false matches. Suppress unnecessary personal fields in the review view.