Home · Business · Customer Service · Feedback & NPS analysis

YES

As of 13 August 2026, AI can analyse comments from your NPS detractors.

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 ita colleague

What the alternative costsNo comparable alternative price is provided in the supplied tool data.

If this goes wrong: you prioritise the wrong customer problem, spend operational effort on it and miss the issue that is actually driving dissatisfaction.

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 detractor comments from your NPS or survey system, including a stable comment identifier, score, date, product or service, and customer segment where those fields are available.
    2. Remove names, email addresses, telephone numbers, account numbers and other details that identify individual customers before sending the data to an AI tool.
    3. Open ChatGPT, Claude or Gemini and paste the survey question, detractor score range, date range, segment definitions and any product context before pasting the anonymised comments.
    4. Paste the copyable prompt and the comments, then ask the tool to produce the requested themes, evidence, identifiers, uncertainties and follow-up actions.
    5. Compare every theme count, proportion and quoted comment against the original export, and correct any comment placed in the wrong theme or counted more than once.
    6. Send the checked analysis to a customer-service or product colleague who understands the survey and ask them to challenge the priorities before you publish a report or change a process.

    Prompt

    Analyse the anonymised NPS detractor comments below. The survey question was: [PASTE QUESTION]. The score range treated as detractor was: [PASTE RANGE]. The comments cover: [PASTE DATE RANGE]. Relevant customer segments or products are: [PASTE CONTEXT].

Do not invent comments, counts, causes, customer details or quotes. Preserve the meaning of each comment and say when a comment is ambiguous. Separate what the comments directly support from your interpretation. Group comments into a small number of non-overlapping themes, explain each theme in plain UK English, give the number and proportion of comments in each theme if the source allows it, and include the source comment identifiers for every theme. Identify sub-themes, repeated phrases, notable differences between segments, and comments that do not fit a theme. Select a few representative quotes only from the supplied text and label them with their identifiers.

Finish with: the three most evidenced problems, the evidence for each, possible customer impact, questions that the data cannot answer, and practical follow-up actions. Do not recommend a business decision solely from the frequency of a theme. Put all uncertainty and data-quality limitations in a separate section. Return the result with headings: Overview, Method, Themes, Segment differences, Unmatched comments, Evidence limits, Priorities for human review, and Suggested next steps.

Comments:
[PASTE ANONYMISED COMMENTS HERE]

    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 a frequent complaint is more important than a less frequent issue affecting a strategically important customer group.
  • AI can merge distinct problems because customers use similar words for different causes.
  • AI cannot establish that the comments represent all detractors or explain the views of customers who left no comment.
  • AI does not own the decision to contact a customer, change a service or allocate resources.
  • AI cannot safely handle identifiable customer data unless your organisation has approved the tool, access controls and processing arrangements.

Even on a YES, the friction has a name: judgement under ambiguity, private data access 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.

AxisScore (0–2)
Output2
Inputs2
Verification1
Liability1
Effort delta2
Total8 / 10

The methodology and its thresholds are published in full.

FAQ

Can AI analyse NPS comments?
Yes. It can group detractor comments, summarise recurring issues, compare segments and link themes back to source comments. You still need to check the evidence and decide which problems matter to your organisation.
How do I use AI to analyse NPS detractor feedback?
Export the comments with identifiers and relevant survey context, remove personal details, then ask the AI to create evidence-linked themes without inventing counts or quotes. Compare its output with the source export before sharing the analysis.
Can AI tell me why my NPS is low?
It can suggest reasons supported by the comments, but it cannot prove why the overall NPS is low. Comments may be unrepresentative, ambiguous or affected by customers who chose not to explain their score.
Is it safe to upload customer comments to AI?
Only use a tool approved by your organisation and remove information that identifies customers unless your data-processing arrangements allow it. Check access, retention and reuse settings before uploading any feedback.

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.