As of 13 August 2026, AI can segment your customer list.

What to actually do

1. Hand it to a person

   The route this page recommends

   A person who owns the outcome does this end to end, worth it when the failure is dear.

2. Use a tool built for this

   Second choice

3. Do it yourself

   The distant third

   A chat interface, chat-fluent skill, and roughly 30 minutes until you can act on the result.

   How to actually do it

   1. Open the customer spreadsheet and create a working copy containing a non-identifying customer ID plus only the fields relevant to the marketing objective.
   2. Remove names, email addresses, phone numbers, full addresses and any sensitive fields that are not necessary, then check that the remaining data was collected and can be used for this purpose.
   3. Write down one segmentation objective, such as retention, repeat purchase or product interest, and define the date range and meaning of each column.
   4. Paste the anonymised data, column definitions and objective into a chatbot using the supplied prompt, or upload the working copy if the interface supports files.
   5. Export or copy the proposed segment assignments and summaries into a spreadsheet, then compare the segment counts and percentages with pivot tables or equivalent spreadsheet calculations.
   6. Inspect a sample of records in every segment against the original rows, correct any missing or contradictory data, and ask the model to rerun the assignments if the data or rules changed.
   7. Name the segments using plain descriptions, select only actions that fit your consent and contact rules, and send the final segment definitions to the colleague who owns the campaign before targeting customers.

   Prompt

   Segment the customer data below for the business objective I provide.
   
   Business objective: [for example, identify groups for retention campaigns]
   Available columns and definitions: [list each column and explain what it means]
   Data:
   [ paste anonymised rows or attach a spreadsheet ]
   
   Use only the supplied fields. Do not infer protected or sensitive characteristics, and do not use names, email addresses, phone numbers or full addresses. If the data is insufficient, say exactly what is missing instead of guessing.
   
   Produce:
   1. A recommended number of segments and the rule or evidence behind that choice.
   2. A table assigning each customer ID to one segment.
   3. A plain-English name and description for each segment.
   4. The number and percentage of customers in each segment.
   5. The key differences between segments, with the source columns that support each difference.
   6. A short suggested marketing action for each segment, clearly labelled as a suggestion rather than a fact.
   7. A list of records that are incomplete, contradictory or difficult to classify.
   8. A validation checklist explaining how I can reproduce the counts and check the assignments in a spreadsheet.
   
   Do not claim that a segment is profitable, loyal, likely to buy or suitable for contact unless the supplied data directly supports that claim.

   Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.