As of 13 August 2026, AI can analyse feedback from your Google reviews.

What to actually do

1. Use a tool built for this

   The route this page recommends

2. Do it yourself

   Second choice

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

   How to actually do it

   1. Open your Google Business Profile and gather the review text, star rating, date and any available review reference into a spreadsheet.
   2. Remove unnecessary personal information from the spreadsheet, keeping only the details needed to link each finding back to the original review.
   3. Add a column for your own notes, such as whether you recognise the incident or whether the complaint has already been resolved.
   4. Paste the spreadsheet rows into a chatbot with the prompt, using separate messages if the dataset is too large for one message.
   5. Ask the chatbot to produce the theme table, supporting review references, reliable counts and proposed priorities without inventing missing information.
   6. Open the original reviews and compare every quoted extract, theme count and urgent finding against the source rows before acting on it.
   7. Choose a small set of verified service actions, then have the relevant colleague check the wording and facts before publishing any public response.

   Prompt

   Analyse the Google review data below for my UK business.
   
   Business type: [describe the business]
   Business goals: [state what you want to improve]
   Review data:
   [paste the reviews or the spreadsheet rows here]
   
   Use only the supplied data. Do not invent review text, themes, counts, customer motives or business facts. Keep positive comments, negative comments and mixed comments separate. Identify recurring themes, give the exact number of reviews for a theme only when it can be counted reliably from the supplied data, and quote short review extracts with their review date or row reference where available. Separate direct customer statements from your interpretation.
   
   Return:
   1. A short overall summary.
   2. The main positive themes.
   3. The main complaints and service failures.
   4. Any recurring issues that appear urgent or reputationally risky.
   5. A table with theme, supporting review references, evidence, confidence and suggested business action.
   6. Which findings need a human to investigate before action.
   7. A short list of practical priorities, ordered by likely customer impact and ease of checking.
   
   Do not diagnose customers, identify private individuals, infer protected characteristics or recommend publishing automatic replies. Finish with questions I should answer before changing policy, training staff or responding publicly.

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

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.