Times of Pakistan

Here’s how Amazon’s AI hosts are now selling you products

2 hours ago 1
ARTICLE AD BOX

When this feature appears on a product page, users can generate an audio discussion.

Amazon has quietly introduced a new AI feature that creates short, podcast-style audio segments where two virtual hosts talk about a product’s reviews, features, and overall value, and even answer user questions like a call-in show.

When this feature appears on a product page, users can generate an audio discussion in which the AI hosts explain key highlights, common complaints, and useful details gathered from different sources. The content is not limited to product listings or customer reviews, as it also includes information from third-party websites and editorial content.

The feature is powered by Amazon Bedrock, which combines multiple AI technologies. Shoppers can also ask questions during the audio session, and the AI hosts respond to them within the same podcast-style format.

A similar concept was introduced earlier by Google through its NotebookLM, which allowed users to turn documents and web pages into conversational audio summaries. While that tool focused on simplifying complex reading material, Amazon’s version applies the idea to online shopping.

The goal is to make it easier for users to understand products that often have long descriptions, technical details, and many customer reviews. However, it is still unclear whether this audio format will truly simplify the shopping experience or just add another layer of information.

Testing has shown some limitations. For example, when asked to compare a mini vacuum with a larger model, the AI could not provide a clear answer and instead suggested checking another product page, something a simple search could have done more quickly.

Amazon has not shared details about which products support this feature. In some cases, trying to generate these audio segments for certain items did not work, suggesting the feature is still being rolled out gradually without clear criteria.

Read Entire Article