As of 13 August 2026, AI can only partly analyse customer feedback about your business on social media.

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 public comments, replies and direct feedback from each social platform for a defined date range, retaining the date, platform, text and any supplied reaction or engagement fields.
   2. Remove names, usernames, contact details, order numbers, message IDs and other personal data that the analysis does not need, then split the material into manageable batches if it is too large to paste at once.
   3. Open a chatbot, paste the prompt and add the anonymised feedback below it, including the date range and platform labels.
   4. Ask the model to analyse each batch using the same theme labels, then ask it to combine the batch results without double-counting repeated comments.
   5. Paste the combined analysis into a spreadsheet and compare every reported count, percentage and quoted comment with the original export.
   6. Open a sample from each theme, including ambiguous and negative comments, and change labels where sarcasm, context or duplicate posts make the model's classification inaccurate.
   7. Send the checked summary and prioritised actions to the colleague responsible for customer service, product or marketing, and agree which actions are supported by evidence rather than by the model's ranking alone.

   Prompt

   Analyse the social media feedback below for my business. The feedback covers [DATE RANGE] and comes from [PLATFORM OR PLATFORMS].
   
   First, remove or ignore names, usernames, email addresses, telephone numbers, order numbers and other personal data that are not needed for analysis. Do not identify or profile individual customers.
   
   Produce:
   1. A short overall summary based only on the supplied feedback.
   2. The main themes, with the number of comments in each theme and the percentage only when the count and total are clear.
   3. Sentiment for each comment as positive, neutral, negative or mixed, with a brief reason.
   4. The most common complaints, praise and requests, keeping separate issues that are not genuinely the same.
   5. A list of representative quotes, copied exactly and anonymised where necessary. Do not invent or rewrite quotes.
   6. Any change over time or difference between platforms, but only where the supplied dates and platform labels support it.
   7. Three practical actions, ordered by likely customer impact and supported by the feedback.
   8. A separate list of uncertain, ambiguous, sarcastic, duplicated or out-of-scope comments.
   
   Do not invent figures, comments, customer motives or trends. Do not treat likes, shares or a single loud comment as proof of importance unless those figures are supplied. State what cannot be concluded from this dataset. Show the calculation behind every count or percentage. Keep the analysis descriptive rather than making decisions about individual customers.
   
   Feedback:
   [PASTE ANONYMISED SOCIAL MEDIA EXPORT OR COMMENTS HERE]

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