Don’t you hate scrolling through hundreds of reviews just to see if that product you’re searching for on Amazon is worth the money? Well, why not sit through a mini-podcast episode talking about that product? Sure, that sounds outlandish, but it just might be a thing. Amazon just unveiled its new AI shopping experts. This feature is basically like Audio Overviews for Amazon.

We have to give props to Google for pioneering this potential new AI trend. Audio Overviews is a feature that originated in Google’s NotebookLM platform. This takes the information that you have in your notebook and converts it into an extremely realistic-sounding podcast episode with two hosts. It was one of the most exciting features announced during last year’s Google I/O. Users can try this feature out now by going to the NotebookLM website or downloading the new NotebookLM app.

Amazon unveiled its AI shopping experts

Amazon’s AI endeavors have been lukewarm to say the least. So, a feature as ambitious as this seems to be right out of left field. Right now, a limited selection of US-based users will be able to try out this feature on the Amazon app. 

If you have this feature, you will see a Hear the highlights button appear on certain product pages. A good example is this Ninja Blender. Tapping on that button will bring up a little audio player, and your personalized micro podcast will begin.

Just like Google’s Audio Overviews, Amazon’s tool will generate a short discussion between two hosts. The discussion will be around two minutes on average and will focus on the product you’re viewing. The hosts will talk about the features, the customer reviews, and who the product is for. Amazon will use information from reviews and the web to inform users about the product.

How they work

The dynamic of Amazon’s offering differs from Google’s in a pretty significant way. Think of Google’s Audio Overviews as a podcast equally hosted by two people. So, each episode, you’ll hear those two people’s voices. On the other hand, even though Amazon’s episodes will involve two people, it’s better to think of it as a podcast hosted by one person who has an additional guest each episode.

Each episode will begin with one person who reminds the user that they are the AI-generated host. A few seconds later, the host introduces their guest. Depending on the type of product being viewed, the guest will actually be a different person. For example, the guest for the blender linked above is a culinary expert named Sophia. However, if you hear an episode for a skin care product, the guest will be a skin care expert named Emma. The audio expert is named Max.

This approach might make Amazon’s AI shopping experts more appealing. It’s because the podcast is introducing the guests as experts in their fields, so it will feel like you’re getting advice from an expert.

As for the actual quality of the voices, it’s a pretty mixed bag. While the voices do sound realistic to a degree, they are still a far cry away from what we hear with Google and OpenAI. Honestly, the voices sound pretty dry and lifeless. Most of the time, it sounds like people are reading from a script rather than having an open discussion about the product. However, we have to give Amazon credit for emulating the host popping the mic whenever they pronounce words with the letter “P.”

Is this necessary?

The main question here is whether or not people would rather sit and listen to a podcast episode about a product. Much of the time, people comb through the page and glance at the reviews to gather information about a product. So, we’re not sure if people would be willing to sit for several minutes listening about an item just to learn about it.

On the other hand, this may not apply to everyone. There are people who absorb information much more easily by listening than by reading. Having all the information boiled down to a short two-minute-long podcast episode might make it easier for certain people to fully understand their product. The episodes pull information from several sources, and they talk about the most notable features. So, this could actually be a make-or-break feature for some people. If anything, this could be a short-lived gimmick that charts the path for more useful features to come in the future.

There’s still the problem with fake reviews

As mentioned before, Amazon will pull information from reviews to help generate the podcast. However, Amazon has a massive issue with fake reviews being plastered on the site. Companies that deliver poor and low-quality devices often flood their products with high reviews to boost them through the rankings. While Amazon does take steps to reduce this, fake reviews still run rampant.

It’s very possible that the Amazon AI shopping experts can draw from these fake reviews and artificially boost the product. If that happens, then that will damage this feature’s reputation. Only time will tell if that happens.