Home · Business · IT, Data & Security · Data analysis & reporting

YES

As of 13 August 2026, AI can find themes in your customer reviews.

Most people should do this themselves; the prompt is on this page.

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 professional analyst or customer-insight specialist is the alternative; no price is stated in the supplied tool data.

If this goes wrong: an important complaint is grouped into the wrong theme and you make a poor business decision until the original reviews expose the mistake.

What to actually do

  1. Do it yourself

    The route this page recommends

    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 reviews from each relevant channel into one spreadsheet, keeping the review text and adding a unique row number, date, rating, product or service, location and channel where available.
    2. Remove names, email addresses, telephone numbers, order numbers and other personal data that the analysis does not need, then check that you have permission to upload the remaining data.
    3. Write down the business decision the analysis is meant to inform, such as improving delivery, fixing a product issue or understanding repeat complaints.
    4. Paste the prompt into a chatbot and attach or paste a manageable batch of reviews, stating the total number of rows and whether the batch is complete.
    5. Ask the model to produce the themes, counts, sentiment and supporting row numbers using only the supplied reviews.
    6. Open the original spreadsheet and sample reviews from every proposed theme, comparing the cited row numbers, quotations and counts with the source text.
    7. Merge duplicate themes, restore important minority issues and change the priority order where your product knowledge shows that frequency alone is misleading.
    8. Send the checked theme table and its evidence to the colleague responsible for the relevant customer or product decision, rather than treating the model's recommendations as findings.

    Prompt

    Analyse the customer reviews below and find the main recurring themes.

Context:
- Business or product: [describe it]
- Review period: [dates]
- Review source: [website, survey or support channel]
- Intended decision: [what you may change as a result]

Instructions:
1. Remove or avoid repeating names, email addresses, telephone numbers, order numbers and other personal data.
2. Group reviews into distinct themes based on what customers are talking about, not just on individual words.
3. Give each theme a short name, a plain-English description, the approximate number of reviews mentioning it, and whether the comments are mainly positive, negative or mixed.
4. For every theme, cite the review IDs or row numbers that support it and include no more than two short quotations from the supplied text.
5. Separate a theme from a possible cause. Mark causes as hypotheses unless the reviews directly support them.
6. Identify important minority issues, contradictory feedback, duplicated reviews and reviews that are too vague to classify.
7. Do not invent counts, quotations, customer details or explanations. If the data is insufficient, say exactly what cannot be concluded.
8. Finish with a table ordered by likely business importance, using evidence from frequency, severity and clarity. Suggest questions I should answer before taking action.

Return the analysis with these headings: Method, Main themes, Minority and contradictory issues, Data limitations, Priorities, Questions to investigate.

Reviews:
[PASTE THE REVIEWS HERE, INCLUDING A UNIQUE ID OR ROW NUMBER FOR EACH REVIEW]

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

  2. Use a tool built for this

    Second choice

  3. Hand it to a person

    The distant third

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

What it gets wrong

  • AI cannot know whether a frequently mentioned issue is commercially important without your product, customer and operational context.
  • AI can merge different problems because they use similar language, or split one problem into several labels because the wording varies.
  • AI cannot establish that the reviews represent all customers when the source contains response bias, duplicate submissions or missing channels.
  • AI cannot safely infer causes from complaints unless your data contains evidence linking the cause to the outcome.
  • AI does not remove your responsibility for protecting personal data or for acting on the analysis.

Even on a YES, the friction has a name: judgement under ambiguity, private data access 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
Liability1
Effort delta2
Total8 / 10

The methodology and its thresholds are published in full.

FAQ

Can AI analyse customer reviews?
Yes. It can group recurring topics, compare positive and negative feedback, and summarise evidence from a supplied set of reviews. Check the proposed themes, counts and quotations against the original rows before acting on them.
How do I use AI to find themes in customer reviews?
Put the reviews in a numbered spreadsheet, remove unnecessary personal data, and give the model the review text with relevant fields such as date, rating and product. Ask for named themes, supporting row numbers, sentiment, minority issues and data limitations, then verify the output against the source.
Can AI tell me what my customers are unhappy about?
It can identify repeated negative topics in the reviews you provide. It cannot prove that those topics represent all customers or explain the underlying cause without better data and your business context.
Is it safe to upload customer reviews to AI?
Only upload data you are permitted to share, and remove names, contact details, order numbers and other unnecessary personal data first. Check the tool's data handling terms and use an approved business account if your organisation requires one.

Nearby answers

All Data analysis & reporting 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.