Let AI clean the customer list you've been avoiding

Every business has The Spreadsheet: 2,000 rows of customers, half of them duplicates, formats from three CRMs ago. Cleaning it is perfect AI work — pattern matching at scale, zero creativity required.

1. Export to CSV. Work on a copy — always.
2. Ask for a diagnosis first: "List the data-quality problems you see, with counts, before changing anything." You'll get: 214 near-duplicates, 3 phone formats, mixed casing…
3. Approve fixes in batches: normalize phones, merge obvious duplicates, standardize company names, flag invalid emails.
4. Demand a change log — every merged row and rewritten value, listed, so you can audit and undo.

Claude's data tools or Claude Code handle this conversationally on files of real size. The result isn't just tidiness: deduped lists cut your email-tool bill and stop the "why did we message them twice" complaints.