As of 13 August 2026, AI can summarise thousands of competitor reviews.

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, power-user skill, and roughly 1 hour until you can act on the result.

   How to actually do it

   1. Open the review sources and export the competitor reviews into a spreadsheet, retaining the review text, rating, source, review ID or URL, competitor, location if available, and publication date.
   2. Remove obvious duplicates, empty rows and unnecessary personal information, then add a note describing what was removed and why.
   3. Create one consistent column for each field and label missing values clearly instead of filling them in, then save the file in a format the AI tool can read.
   4. Write down the competitors, UK market scope, business decision and themes you want investigated before uploading the file.
   5. Upload the dataset to Julius AI or a chatbot, paste the prompt, and ask it to produce the comparison tables with review IDs, counts and quotations tied to the source rows.
   6. Check every reported count against spreadsheet filters or pivot tables, and open a sample of source reviews for each major theme to confirm that the wording and sentiment match.
   7. Rewrite the strategic implications so they distinguish review evidence from your own judgement, then send the checked summary to the product or strategy decision-makers.

   Prompt

   Analyse the competitor review dataset below for the UK market. Do not invent facts, reviews, themes, counts, quotations or causes. Treat each row as one review unless the data clearly identifies duplicates.
   
   Dataset:
   [PASTE THE REVIEW TABLE OR UPLOAD THE FILE]
   
   Known fields:
   [LIST THE COLUMN NAMES]
   
   Competitors to compare:
   [LIST COMPETITOR NAMES]
   
   Business question:
   [STATE WHAT DECISION THIS analysis will inform]
   
   Produce:
   1. A short executive summary separating observations from interpretations.
   2. A comparison table for each competitor covering recurring praise, recurring complaints, unmet needs, product or service expectations, and notable differences.
   3. For every theme, give the review count, share of valid reviews if calculable, competitors affected, rating pattern, and source review IDs. If a count or share cannot be calculated reliably, say so.
   4. Separate themes supported by many reviews from isolated comments.
   5. Identify possible sampling bias, duplicate reviews, missing fields, language issues and other data limitations.
   6. Include short verbatim quotations only when they are present in the dataset, preserving the source review ID and wording. Do not include unnecessary personal information.
   7. End with three evidence-based implications for our strategy and three questions that the dataset cannot answer.
   
   Use plain British English. Do not treat star ratings as sentiment without explaining the limitation. Do not claim that a theme is a market-wide fact when the dataset only supports a statement about these reviews. Show your working assumptions and provide a verification checklist so I can compare the conclusions with the source data.

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