Templates / Data cleaning prompt (plan first, then fix)

Data cleaning prompt (plan first, then fix)

Profile a messy list, propose fixes for your approval, then clean it with a change log, without deleting rows or inventing values.

General AIOpen, no sign-up

Who it is for

Anyone with a messy customer list, product sheet or export who needs it tidy before importing it somewhere.

How to use it

  1. Paste a sample (or the whole list if it is small) and say what the data will be used for.
  2. Read the plan and approve or edit each proposed fix.
  3. Check the before-and-after examples, then spot-check the cleaned data.
  4. For hundreds of rows, ask for a script or spreadsheet formulas instead of pasting everything.

The template

You help me clean a data table. Work in two stages and wait for my approval between them.

Purpose of this data: [WHAT IT WILL BE USED FOR]
Columns that matter most: [LIST]
Data (CSV or pasted rows): [PASTE]

Stage 1, Plan only. Do not change the data yet. Tell me:
1. What you see: columns, apparent types, number of rows, how many are blank per column.
2. Problems: duplicates, inconsistent formats (dates, phone numbers, capitalization, states or countries), obvious typos, columns that mix two kinds of value, and rows that look wrong.
3. Proposed fixes, as a numbered list. For each, give a before and after example from my data.
4. Questions where a fix depends on a choice I must make.
Then stop and ask me which fixes to apply.

Stage 2, After I answer. Apply only the approved fixes and return:
- The cleaned table.
- A change log: what was changed, with up to 10 examples per type of change.
- A list of rows you were unsure about, left unchanged.

Rules:
- Never delete rows. Mark suspected duplicates or bad rows in a new column instead.
- Never invent or guess missing values. Leave them blank.
- Keep the original value in a separate column whenever you change a column.
- Do not merge people or companies unless I approve it.
- If the data is too long for you to process reliably, say so and offer a script or spreadsheet formulas instead.

Example

An illustrative example written by BotPickr for a fictional business. It is not a real model run, so your own output will differ.

Example input

Purpose: import into a newsletter tool. Data:
name,email,phone,signup
jane DOE,Jane@Example.com,(555) 010-0100,10/02/2026
John Smith,john@example,555.010.0101,2026-10-03
jane doe,jane@example.com,5550100100,Oct 2 2026

Example output

Stage 1 plan: 3 rows, 4 columns. Problems: (1) rows 1 and 3 look like the same person; (2) names have inconsistent capitalization (jane DOE to Jane Doe); (3) email in row 2 looks incomplete (john@example has no domain ending); (4) phone formats differ; (5) dates use three formats, and 10/02/2026 could be Oct 2 or Feb 10. Question: are dates month first? Which fixes should I apply?

Guardrails and human approval

These are practical notes, not legal advice. Templates are provided as is. See section 4 of the Terms.

Works with

We have not benchmarked output quality across models. Output varies by model and by what you paste in, so read it before you use it.

Related templates

Small-business automation auditWeekly KPI summary promptPrompt evaluation checklistAll templates

The BotPickr Weekly

Free: 30 copy-paste AI templates, including our “ask before you act” safety prompt, plus new bot picks, in plain English.

No spam, unsubscribe anytime. By submitting, you agree to our Terms and Privacy Policy.