Apify scrapes reviews of your products and your competitors', ChatGPT tags each one, and Airtable shows which complaints keep coming up.
Reviews on Amazon and Trustpilot tell you exactly what buyers dislike about your products and your competitors', but nobody has time to read thousands of them. Teams skim the one-star reviews, remember the angriest ones and miss the quieter patterns. Product and listing decisions end up based on anecdotes.
List 5 to 10 ASINs or Trustpilot domains: your top sellers and the two or three products shoppers compare them with. Keep them in an Airtable table called Sources with a Platform field.
Pick an Amazon reviews scraper and a Trustpilot scraper from the Apify Store. Cap each run at 200 of the most recent reviews per product and create a weekly schedule for Monday morning.
Add an Apify webhook on the ACTOR.RUN.SUCCEEDED event that calls a short script. The script reads the run dataset, sends reviews to the OpenAI API in batches of 20 with the tagging prompt, and requests JSON output.
Create a Reviews table with Product, Rating, Date, Text, Topic (single select), Sentiment, Issue and Review URL. Write records with performUpsert using Review URL as the merge field so re-scraped reviews are not duplicated.
Add an Airtable Interface with a bar chart of Issue counts for the last 30 days, filtered by Product, and a second chart comparing your top issues with competitors'.
Run the summary prompt on the last 7 days of negative reviews and share the output with the product and listings teams before their Tuesday planning meeting.
You will receive a JSON array of product reviews. For each review, return an object with:
- id: copied from the input
- topic: one of Quality, Sizing, Shipping, Packaging, Price, Instructions, Durability, Customer service, Other
- sentiment: positive, neutral or negative
- issue: a specific problem in 3 to 8 words, for example "zipper breaks within a month", or null if none
- feature_request: something the reviewer wishes the product had, or null
Base every field only on the review text. Return a JSON array and nothing else.
Reviews:
{{reviews_json}}Here are last week's negative reviews for {{product_name}} and its competitors, already tagged with topic and issue: {{tagged_reviews}}
Identify the five most frequent issues for our product and the five most frequent for competitors. For each, give the count, one representative quote, and whether it is rising or falling compared with the previous four weeks ({{prior_counts}}).
Finish with three concrete changes to the listing or the product itself that would address our top complaints.Give it a keyword and get a brief built from the pages that rank today: headings, gaps, questions to answer and a target length.
A CrewAI crew checks 50 competitor pages, changelogs and news sources every week and sends a short brief covering only what changed.
Scrape competitor product pages every night with Apify, match them to your SKUs, and update Shopify prices within margin and change limits you set.