Amazon Juices Shopping With Generative AI Summary of Reviews
Amazon’s considerable data science muscle is flexing artificial intelligence in a new corner of e-commerce, with generative AI summarizing shoppers’ product reviews, the company announced Monday.
Armed with the goal of improving the shopping experience by eliminating the need to trawl through one review after another, the e-commerce juggernaut reportedly has been working on the capability for months. At least part of the framework was already there, with a key word cloud of frequently used words. Now, the tech will focus on recurring themes and auto-generate a blurb on the product detail page.
More from WWD
According to the company’s blog, Amazon has been refining its approach to reviews for years, from offering more visibility across borders to filters that conjure only relevant feedback — for instance, the system can focus only on apparel reviews written by someone matching the shopper’s height and weight.
Now “we want to make it even easier for customers to understand the common themes across reviews,” Amazon posted, “and with the recent advancements in generative AI, we believe we have the technical means to address this long-standing customer need.”
It sounds like a fairly straightforward change, but with nearly 1.5 billion reviews flooding the marketplace last year, even incremental updates can create a meaningful impact to its bottom line.
Of course, a lot depends on details, like how the changes were executed or the feature’s scope of development. For instance, it’s not clear how well this system can keep fake reviews from contaminating the summaries, or out-of-date reviews when the seller switches out the product on its sell page — a practice that, apparently, still dogs the platform across numerous product pages that WWD checked, as of Monday.
Amazon nodded to that in its blog. “We continue to invest significant resources to proactively stop fake reviews,” the company wrote. “This includes machine learning models that analyze thousands of data points to detect risk, including relations to other accounts, sign-in activity, review history and other indications of unusual behavior, as well as expert investigators that use sophisticated fraud-detection tools to analyze and prevent fake reviews from ever appearing in our store.”
Amazon has been deploying AI and machine learning in this game of whack-a-mole to eliminate phony products, reviews and sellers for years, and at least on the review side, plenty still get by the system. But as another safeguard, the company said, the system will only use reviews from verified purchases.
The machine-written summaries are rolling out on mobile to consumers in the U.S. for select categories. In short order, within months, the company hopes to expand the feature to more users and product types.
Best of WWD
