Use AI to Plan a Spreadsheet Data Review

A repeated customer name is not automatically a duplicate record. Use AI to help describe review rules and exceptions, then apply those rules to a preserved working copy. The objective is an explainable dataset, not the smallest possible row count.

Prepare the information that matters

Keep the original export unchanged. Describe what one row represents, the expected comparison key and which person owns ambiguous business rules. Share a sanitized sample instead of a customer export when the structure alone is sufficient.

Use a service your organization permits for this material. A pasted excerpt is the assistant's source for this workflow; the prompt does not give it access to files, mail or meetings you have not supplied.

Copy the ready-to-use prompt

Help me plan a careful review of the spreadsheet sample below. First establish what one row represents and which columns identify a record. Ask me if that definition is missing.
Separate identical rows, conflicting records with the same key, legitimate repeated values, blanks and format inconsistencies. Do not assume a repeated name is a duplicate record.
Return proposed review rules, an exceptions table and before/after checks. Keep the source unchanged. Do not delete records, guess which conflicting value is correct or claim to have cleaned a workbook. List the decisions that require the data owner.

SAMPLE AND CONCERNS
[Paste a small anonymous sample and describe what you want to check here]

Work through a concrete example

101/P1: exact-repeat candidate; review the two identical rows.
102/P2: key conflict; quantities differ, so do not delete one as an exact duplicate.
Morgan: repeated customer name; not itself a duplicate key.
Before changing anything: record 4 source rows, define the retention rule and resolve the quantity conflict with the owner.

Why these choices matter

The example separates a repeated row from a disagreement about the value. It does not decide whether to keep the first or last conflicting entry. That decision depends on how the source system creates order lines and what the data owner confirms.

Review and improve the first answer

Challenge the proposed key using the sample. Identify legitimate repeated values that must remain. Distinguish identical rows from conflicting records with the same key. List the decisions that require the data owner.

Ask for a specific correction when the response misses the goal. Changing the requested format can improve readability; it cannot make an unsupported fact true. Return to your source when the answer changes a number, date, responsibility or condition.

Check the result before using it

1. Confirm what a row represents before accepting a key.
2. Keep row counts and known totals from the source copy.
3. Record conflicts separately from exact-repeat candidates.
4. Compare reviewed output with the untouched source and explain every exclusion.

Fix the common failure patterns

Use the plan alongside the preserved source and an exceptions record. A smaller output is not proof of better data; the record owner must be able to explain which rows changed and why.

Find another task in the AI at Work prompt library, or continue with the related DeskPlain guides below.

Original DeskPlain prompts and fictional worked examples, prepared with AI assistance. Examples illustrate the intended reasoning; they are not captured model responses or a claim of testing across AI products. Read the editorial policy.