As of 13 August 2026, AI can compare feedback from different customer segments.

This still needs a person who signs their name to it.

Can you do it?

15 minutesto a draft.

30 minutesto something you’d act on.

Cost, all in£0

Skill neededpower-user

Who has to check ita colleague

What the alternative costsAn alternative is Botpress, an open platform for building LLM chatbots and agents.

If this goes wrong: you act on a misleading segment comparison and direct service or product changes away from customers whose feedback was poorly represented.

What to actually do

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.
Use a tool built for this
Second choice

Do it yourself

The distant third

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

How to actually do it

Open the feedback export and remove names, email addresses, order numbers, free-text identifiers and other personal data that is not needed for the comparison.
Gather the exact segment definitions, survey questions, response scales, NPS calculation rules, collection channels and periods covered by the data.
Check that each row has the correct segment label and that scores, dates and missing responses use consistent formats before uploading or pasting the data into a chatbot.
Paste the supplied prompt, then add the cleaned feedback, segment definitions, scoring rules and business context in the marked sections.
Ask the chatbot to produce the segment table, theme comparison, supporting quotes, data limitations and cautious actions requested in the prompt.
Compare every reported count, average, NPS result and percentage with the original export, and trace each quoted comment back to its source row.
Ask a colleague who understands the customers and data collection process to challenge the segment definitions, theme labels, missing-data warnings and proposed actions before sharing the analysis.

Prompt

Compare the customer feedback below across the named segments.

Segment definitions:
[PASTE THE EXACT RULES USED TO ASSIGN EACH CUSTOMER TO A SEGMENT]

Survey questions and scoring rules:
[PASTE THE QUESTIONS, RESPONSE SCALE, NPS RULES OR OTHER METRICS]

Feedback data:
[PASTE OR UPLOAD THE DATA, INCLUDING SEGMENT, DATE OR PERIOD, SCORE, QUESTION, COMMENT AND ANY OTHER RELEVANT COLUMNS]

Business context:
[DESCRIBE THE PRODUCT, SERVICE, CUSTOMER JOURNEY AND DECISIONS THIS ANALYSIS MAY INFORM]

Produce:
1. A table showing the number of responses and the available score metrics for each segment.
2. The main positive and negative themes in each segment, with the number of supporting comments where the data allows it.
3. Themes shared across segments and themes that appear distinctive to one segment.
4. Representative quotes for each important finding, labelled with their segment, without exposing names, email addresses, order numbers or other personal data.
5. Differences that may be caused by unequal sample sizes, missing data, question wording, collection method or time period.
6. A short list of cautious, testable actions, separating evidence from interpretation.

Do not invent data, quotes, sample sizes, statistical tests or causes. Do not claim that one segment caused an outcome. State when a comparison cannot be supported by the supplied data, show the calculation for every reported metric, and flag findings that need human review. Keep the language plain and distinguish clearly between counts, calculated differences and interpretation.

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

What it gets wrong

AI cannot know whether your segment definitions reflect meaningful customer differences or accidental differences in how data was collected.
AI can turn a small or biased set of comments into a convincing theme without establishing that the theme represents the wider segment.
AI cannot establish that a segment caused a lower score or that a proposed service change will improve the result.
AI does not carry responsibility for decisions that exclude, deprioritise or otherwise disadvantage a customer segment.
AI cannot replace a colleague who knows the history behind unusual feedback, product changes and customer relationships.

Even on a YES, the friction has a name: context depth, judgement under ambiguity and verification cost.

How we scored this

Five axes, each scored nought to two by hand: ten means AI carries the task cleanly, and the thresholds that turn a total into YES, PARTLY or NO are published in the methodology. Each axis name links to its definition.

Axis	Score (0–2)
Output	2
Inputs	2
Verification	1
Liability	1
Effort delta	2
Total	8 / 10

The methodology and its thresholds are published in full.

FAQ

Can AI compare customer feedback by segment?: Yes. It can compare scores, group comments into themes, identify shared and distinctive issues, and link findings to representative quotes. Give it the segment definitions and sample counts, then check the calculations and interpretations against the source data.
Can AI analyse NPS by customer segment?: Yes, if you provide the underlying scores and the exact NPS calculation rule. It can calculate and compare the results, but you still need to check response counts, missing data, collection methods and whether the differences are meaningful.
Can AI tell me which customer segment is most unhappy?: It can identify the segment with the lowest supplied score or the most negative coded feedback. That does not prove the segment is broadly the most unhappy, because response rates, sample sizes, question wording and comment selection can distort the comparison.
Is it safe to upload customer feedback to AI?: Only upload data your organisation permits you to use and remove personal data that is not needed for the analysis. Check the chatbot provider's data handling terms and use an approved business account where your organisation requires one.

Nearby answers

All Feedback & NPS analysis rulings

Assessed by gpt-5.6-luna (gpt-5.6-luna) on 2026-08-13, second-checked by an independent model. Wrong somewhere? Email [email protected] and it gets re-checked.

The newsletter

AI news, new answers and product picks, straight to your inbox.