A Chinese AI blogger had Alibaba QwenWork gather listings, prices and reviews for a product category across several shopping platforms and analyze why customers complain. A friend who does e-commerce sentiment analysis said a report that used to take half a day or more came out in one run.
- First reported
- 163.com · 28 Jul 2026
- Added
- 13 Oct 2026
- Source checked
- 13 Oct 2026
What they did
No clean spreadsheet was prepared in advance. QwenWork used an e-commerce data source to search several platforms, saved the products, prices and reviews to a CSV, then analyzed negative-review causes, customer sentiment and business issues, with charts on trends, risks, platforms and customer sources.
What happened
The run took about 30 minutes. The friend said it separated positive and negative reviews, grouped the problems and marked trends, and would be easier to explain to a product team than an API classification script.
Keep in mind
The product category is not named and the article shows no accuracy check on the scraped data. It was tested during the closed beta and the run was slow.
Try it with your agent
Find <product type> listings on <platform 1>, <platform 2> and <platform 3>. Save product name, price and customer reviews to a CSV. Then group the negative reviews by cause, show how sentiment changed over the last <number> weeks, and list the three problems we should fix first, quoting example reviews.
Replace the <angle brackets> with your own details.
Original source
163.comUser storySomeone who uses the agent described their own setup. Figures are theirs.
Tools for this kind of work
- RepoExtract tables and text from complex or scanned PDFs with Docling
- RepoSearch the web and scrape sites into clean data with Firecrawl
