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

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 1 hour until you can act on the result.

   How to actually do it

   1. Open your customer database or spreadsheet and export a cleaned table with one row per customer, removing direct identifiers and keeping only fields relevant to the business purpose.
   2. Write down the decision the segments will support, such as choosing customer communications, prioritising support or understanding purchase patterns, and note the date range covered by the data.
   3. Check the export for duplicate customer rows, missing values, inconsistent categories and impossible dates before sending it to an AI tool.
   4. Open a permitted chatbot or a data-analysis tool such as Akkio, paste the prompt, then upload or paste the cleaned table and the list of fields.
   5. Ask the tool to produce the segment definitions, record counts, percentages, customer-ID mapping and reproducible calculations requested in the prompt.
   6. Compare every count, percentage and segment assignment against the source spreadsheet using filters, pivot tables or formulas, and correct any mismatch before using the result.
   7. Ask a colleague who understands your customers to challenge whether the segments are useful, fair and consistent with the stated business purpose.
   8. Save the approved segment definitions and mapping in your reporting system, and restrict access to the customer-level data to people who need it.

   Prompt

   I need to segment customer data for this business purpose: [describe the decision or campaign].
   
   I will provide a table containing these fields: [list fields]. The data has been cleaned as follows: [describe cleaning, missing values and date range]. Use customer IDs only and do not reproduce names, email addresses, phone numbers or other direct identifiers.
   
   Create a practical segmentation analysis. First identify data-quality problems, duplicate records, missing values and fields that should not be used. Do not infer or use sensitive or protected characteristics, and do not claim that a segment is causal. Compare several defensible segmentation approaches, explain which one you recommend and why, then produce:
   
   1. A clear name and plain-English description for each segment.
   2. The rules or measurable characteristics that place a record in each segment.
   3. The number and percentage of records in each segment.
   4. The main differences between segments using only fields present in the data.
   5. A table mapping each customer ID to its segment.
   6. Limitations, possible bias, and checks I should perform before using the result.
   7. A short list of appropriate business actions for each segment, clearly labelled as suggestions rather than facts.
   
   Show your calculations or formulas so I can reproduce them in a spreadsheet. If the data or business purpose is insufficient, ask focused questions instead of guessing. Do not invent values, customer behaviour, causes or recommendations that are unsupported by the data.

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