How to Clean a Messy Customer List With AI Before Your Holiday Emails Go Out

Open wooden index card box with neatly stacked blank cards and a eucalyptus sprig on a mint green background

7 min read

Have you ever sent a holiday email that opened with “Hi JENNIFER,” or bounced off an address with a stray space in it? If your customer list lives in a spreadsheet that grew one copy and paste at a time, the answer is probably yes. The good news: cleaning a messy list is one of the most reliable, low-risk jobs you can hand to AI, and you can do it in under an hour this week, before the busiest sending season of the year.

This guide walks you through it step by step. You will see how to do it with the Clean Data feature inside Microsoft Excel’s Copilot, and how to get the same result with a general chat assistant like ChatGPT, Claude or Gemini if you do not pay for Copilot. No formulas required.

Why a tidy list matters more in November than any other month

A customer list is only as useful as its messiest column. Duplicate rows mean the same person gets two copies of your sale email (and a reason to unsubscribe). Inconsistent city or state names break any filter you try to build, so “send only to my San Diego customers” quietly skips everyone typed as “san diego ” with a trailing space. Numbers stored as text make your “total spent” column add up wrong.

None of these problems are hard to fix. They are just tedious, which is exactly the kind of work AI is good at. And with shoppers expected to lean on AI shopping tools more than ever this season (we covered that in our look at where a solo seller should spend the weeks before Thanksgiving), the owners who reach their existing customers cleanly will have an edge.

Step 1: Make a safety copy before anything touches your data

Before you let any tool change a single cell, save a copy of the file and name it with today’s date, for example “customers-original-2026-11-10”. This takes ten seconds and means every later step is reversible. If an AI suggestion turns out to be wrong, you simply go back.

While you are at it, look at the columns you actually use. A typical solo business list has name, email, phone, city, last purchase date and total spent. If you have columns you never use, hide them for now. Less clutter means clearer suggestions.

Step 2: Turn your list into a proper table

Both Excel and AI chat tools work best when your data is a clean grid: one header row, one customer per row, no merged cells, no notes typed in the margins. In Excel, click anywhere in your data and choose Home, then Format as Table. Microsoft’s own help page notes that Clean Data works best when your data is formatted as a table.

Check for these quick structural problems by eye before moving on:

  • Blank rows splitting the list into chunks
  • Two header rows (often left over from an old export)
  • A “notes” column that mixes phone numbers, birthdays and reminders in one cell

Step 3, option A: Use Clean Data with Copilot in Excel

If your Microsoft 365 plan includes Copilot, Excel has a built-in cleaner. According to Microsoft’s support guide for Clean Data, you select your data, open the Data tab, and choose Clean Data. A side pane then shows suggestion cards for three kinds of problems:

  • Inconsistent text, such as the same value typed with different capitalization or punctuation
  • Inconsistent number formats, where some cells in a number column are stored as text
  • Extra spaces at the start, end or middle of values

Microsoft’s Excel team described the newer version of this feature in its October post, “Beyond analysis: Copilot gets to work in Excel”, saying Copilot can spot potential issues, explain what it found and recommend corrections before you start any analysis. Their example will look familiar: one column containing “USA”, “US” and “United States” all at once.

A few honest notes so you are not surprised. Microsoft says the feature is rolling out gradually, so it may not appear in every account or region yet. It performs best in English. And the original announcement suggested it works best on tables up to roughly 100 columns and 50,000 rows, which is far bigger than most solo business lists, so size is unlikely to be your problem.

Review each card before you accept it. Treat the suggestions like a helpful assistant’s to-do list, not like a command. If a card wants to merge “Smith Co” and “Smith Co.” that is probably right. If it wants to merge two different customers who happen to share a surname, say no.

Step 3, option B: Use any AI chat assistant instead

No Copilot? You can get most of the same benefit from a general assistant. Here is the safe way to do it.

  1. Strip what the AI does not need. For a cleanup job, the assistant rarely needs full email addresses or phone numbers to fix capitalization and city names. Make a working copy that keeps only the columns you want cleaned, plus a simple row number so you can match results back.
  2. Export that copy as a CSV file (File, then Save As, then CSV).
  3. Upload it with a precise instruction. Try: “This is my customer list. Do not delete any rows. Standardize capitalization of names and cities to Title Case, remove extra spaces, make every state a two-letter code, and flag likely duplicates in a new column called Possible_Duplicate instead of removing them. Return a CSV and a short list of every change type you made.”
  4. Spot-check ten random rows against your original file before you trust the rest.

The phrase “flag, do not delete” is the most important part of that prompt. AI assistants are confident, and a confident tool that deletes rows is how real customers disappear from your list. For a second opinion, ask the assistant to list the five changes it is least sure about.

Step 4: Decide duplicates yourself

Duplicates are where judgment matters. “Maria Lopez, maria@…” and “Maria López, mlopez@…” might be one customer with two emails, or two different people. AI can find candidates in seconds, but only you know your customers.

A simple rule that works for most solo businesses: if two rows share an email address, merge them. If they share only a name, keep both and add a note. When you merge, keep the most recent purchase date and add the totals together.

Step 5: Lock in the habit so the mess does not come back

A clean list stays clean only if new entries arrive clean. Three small habits help:

  • Use a form (Google Forms, Microsoft Forms, or your website’s signup form) instead of typing contacts by hand, so formats stay consistent.
  • Use drop-down lists in Excel or Google Sheets for fields like state or customer type.
  • Run a five-minute cleanup on the first of each month, using the same prompt you saved from Step 3.

If you are starting to wonder whether a spreadsheet is still the right home for your customers at all, our side-by-side of spreadsheets, free CRMs and AI assistants for tracking leads will help you decide. And if you want the cleanup to run on its own, an automation tool can trigger it; see what one AI step really costs in Zapier, Make and n8n first.

Your 45-minute plan for this week

  1. Save a dated copy of your list (1 minute).
  2. Format it as a table and remove structural clutter (10 minutes).
  3. Run Clean Data in Excel, or the CSV prompt in your chat assistant (15 minutes).
  4. Review duplicates by hand using the email rule (15 minutes).
  5. Save your prompt and set a monthly reminder (4 minutes).

Do it once and your holiday emails will reach the right people, with the right names, the first time. Which column in your list is the messiest right now? Start there.

Related reading

Leave a Comment

Scroll to Top