Customer Review Analyser — Turn 100 Reviews Into Actionable Product Insights in 60 Seconds
Built by voidrunner ·
About this build
A Claude powered workflow that scrapes or accepts pasted customer reviews and outputs a structured insight report covering what customers love, what they complain about, hidden feature requests, and the exact words customers use to describe the product.
What was built
The idea came from spending way too long manually going through reviews for a product launch. I kept copy pasting reviews into a doc, highlighting things, losing track of what was a one-off complaint versus a pattern. It was messy and I kept second guessing whether I had missed something. The workflow is simple. You paste in a batch of reviews, anywhere from 20 to 200, and the prompt does five things in one pass: Identifies the top 5 things customers consistently praise with the actual phrases they use, not paraphrased. Lists the top 5 recurring complaints ranked by how often they appear. Surfaces any feature requests or suggestions that keep coming up even if customers frame them as complaints. Pulls out the most emotionally charged phrases, positive and negative, that could be used directly in marketing or addressed in support responses. Gives an overall sentiment summary with a rough percentage split between positive, neutral and negative. The phrase extraction part is the most useful piece. Instead of summarising what customers said in my words, it gives me the actual language real buyers use, which is gold for writing product descriptions, ads and email subject lines. Tested this on three different product review sets. Took about 4 seconds per run. Caught two recurring complaints I had completely missed when reading manually, one about packaging and one about sizing information being unclear on the listing. Happy to share the full prompt in the comments if anyone wants to test it on their own review set.