Use AI to analyze customer reviews and uncover hidden patterns, sentiment trends, and actionable insights.
Tools: ChatGPT, Gemini
Time required: 45 minutes
Step-by-step instructions
- Collect your reviews: Copy reviews from Google, Yelp, Facebook, or wherever customers leave feedback
- Paste all reviews into a Google Sheet (one review per row) or a single document
- Go to ChatGPT or Gemini and use the prompt below
- Copy all your reviews into ChatGPT (you can paste up to 50 at once) or ask to analyze the google sheet with embedded Gemini if you have a paid subscription
- Review the analysis: patterns, sentiment breakdown, and specific quotes
- Create an action plan addressing the top 3 issues or opportunities identified
Example prompt
Analyze these customer reviews for my [type of business]. For each category below, provide specific examples with direct quotes:- SENTIMENT BREAKDOWN: What percentage are positive, neutral, negative?
- TOP PRAISE THEMES: What do customers love most? (List top 5)
- TOP COMPLAINT THEMES: What frustrates customers? (List top 5)
- HIDDEN OPPORTUNITIES: What do customers want that we don't currently offer?
- COMPETITIVE MENTIONS: Do customers compare us to competitors? What do they say?
- WORD CLOUD: What words/phrases appear most often?
End with 3 specific, actionable recommendations based on this analysis.[PASTE REVIEWS HERE]
Business benefit: Kenko Tea used AI sentiment analysis to detect subtle emotions in customer reviews and discovered packaging complaints about their "inconvenient resealable pouch." After redesigning based on this insight, they reduced negative reviews by 50% and increased customer satisfaction scores by 10%. Source: HubSpot
Step-by-Step Guide
1
Export or copy 30-100 recent reviews from Google/Yelp/social into one doc or sheet (one per row).
2
Clean the data: include date, rating, channel, and remove duplicates or obvious spam.
3
Paste into ChatGPT/Gemini with the prompt, adding your business type and audience.
4
Ask the model to cluster themes and pull verbatim quotes per theme with sentiment tags.
5
Review the 5 recommended actions and choose the top 3; assign owner and deadline.
6
Draft response templates for common complaints and publish one quick win this week.
Pro Prompt — Copy & Use
You are a CX/research analyst. Analyze these customer reviews for [business] serving [audience] across [channels/locations].
Output:
- Sentiment split (%) with confidence and top adjectives per polarity
- Thematic clusters (name each theme) with 2-3 verbatim quotes and frequency
- Praises vs complaints vs latent desires (what they wish existed)
- Competitor or alternative mentions + why they prefer/avoid them
- Word/phrase spikes (top 15) and emerging trends over time if dates exist
- 5 prioritized actions (impact/effort matrix) with suggested owner/metric to watch
- 3 messaging angles to amplify what people love
Flag data gaps or anomalies and suggest 3 follow-up data pulls.Recommended Tools
Pro Tip
Kenko Tea: Negative reviews about packaging dropped by half; customer satisfaction scores increased by 10%. Source: HubSpot
Your Challenge
Export your last 30 customer reviews and run them through ChatGPT. Identify one quick win you can implement this week.
