Home · Business · Customer Service · Feedback & NPS analysis

YES

As of 13 August 2026, AI can find out why customers would not recommend your business.

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

Can you do it?

5 minutesto a draft.

30 minutesto something you’d act on.

Cost, all in£0

Skill neededchat-fluent

Who has to check ityou

What the alternative costsA human analyst or survey specialist is the alternative; no price is supplied here.

If this goes wrong: you fix the wrong customer problem, spend money on the wrong change and overlook a smaller but more serious complaint.

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. Export the relevant recommendation scores, open-text answers and complaint notes from your survey or customer-service system, and remove names, contact details, order numbers and other identifiers.
    2. Open a chatbot and paste a short description of your business, the customer group, the survey period, the recommendation question and the meaning of any score labels.
    3. Paste the anonymised feedback beneath the prompt, keeping each response on a separate line with its score and a simple respondent ID such as R001.
    4. Ask the chatbot to produce the theme table, supporting comment counts and short anonymised evidence quotes required by the prompt.
    5. Compare every reported theme and count with the original export, checking that positive, neutral and negative comments have not been merged or omitted.
    6. Ask a colleague who knows the service to challenge the three proposed reasons and identify any important operational context missing from the data.
    7. Turn the surviving reasons into a small set of follow-up questions or service checks, then investigate those before committing budget or changing the customer experience.

    Prompt

    Analyse the anonymised customer feedback below to find out why customers would not recommend this business.

Business and service context:
[PASTE A SHORT DESCRIPTION OF THE BUSINESS, CUSTOMER TYPE, PRODUCTS OR SERVICES, LOCATIONS AND RELEVANT PERIOD]

Feedback data:
[PASTE CUSTOMER RECOMMENDATION SCORES, OPEN-TEXT ANSWERS AND RELEVANT COMPLAINT OR SUPPORT NOTES]

Do not invent facts, motives, customer segments or causes. Treat a comment as evidence of one customer's experience, not proof of a general trend. Separate direct statements from your interpretation. Group comments into clear themes and include the number of comments supporting each theme, using only the supplied data. Quote short anonymised excerpts as evidence, removing names, contact details, order numbers and other identifying information. Distinguish reasons for dissatisfaction from reasons specifically linked to not recommending the business. Flag contradictory feedback, unclear comments, possible duplicates and themes supported by only one comment. Compare detractors with neutral or positive respondents if those responses are supplied. End with: (1) the three best-supported reasons customers would not recommend the business, (2) the evidence for each, (3) what additional question or data would test each explanation, and (4) practical actions to investigate before making a major change. Do not claim that the feedback proves causation.

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

What it gets wrong

  • AI cannot tell whether the respondents represent all of your customers or only the people motivated to answer.
  • AI cannot prove that a complaint caused a low recommendation score rather than merely appearing alongside it.
  • AI cannot supply missing operational context, such as a recent outage, staffing change or policy decision, unless you provide it.
  • AI cannot decide which customer problem deserves investment when the evidence is mixed and the trade-offs are commercial.
  • AI can miss sarcasm, coded language, duplicate responses and complaints expressed without the words you expected.

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

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.

AxisScore (0–2)
Output2
Inputs2
Verification1
Liability2
Effort delta2
Total9 / 10

The methodology and its thresholds are published in full.

FAQ

Can AI analyse my NPS detractors?
Yes. It can group detractor comments, count supporting responses and show short evidence quotes, provided you supply the underlying feedback. Check the themes against the original responses because grouping is an interpretation, not proof of cause.
Can AI tell me why customers are unhappy?
It can identify recurring reasons stated in the feedback, such as service delays or product problems. It cannot reliably discover reasons customers did not mention, or prove that one issue caused their dissatisfaction.
Can ChatGPT analyse customer survey responses?
Yes, for a first-pass analysis of anonymised survey responses. Give it the question wording, score definitions and business context, ask it to separate evidence from interpretation, and check its themes against the raw export.
How do I find out why customers would not recommend my business?
Start with recommendation scores and written comments, then compare detractors with neutral and positive respondents. Use AI to organise the evidence, but test its proposed reasons with follow-up questions, complaint records and people who understand the service.

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.