AI in the wild: What we can learn from Amazon & AllTrails

AI in the wild: What we can learn from Amazon & AllTrails

Listen to this article

The internet is humming with excitement. The rising popularity of Artificial Intelligence (AI), specifically large language models (LLMs), has spurred a wave of new products and generated immense returns for investors.

According to a recent report by McKinsey & Co., corporate adoption of AI has reached upwards of 55%.

Yet, the question remains: how valuable will AI be going forward? For product managers, and anyone involved in building a digital product, the question is, ‘How do I use AI to build a better product?’

The temptation is to rush to integrate AI in some way, any way, and hope that your users like it. Even Google has fallen victim to this trap, though they were under immense pressure because of their position in the market.

Integrating AI requires a deep understanding of the customer journey and possible points of friction along the way.

The question that you can ask is ‘How much am I asking of my user, and how can I ask less?’

Two examples of this thinking in practice are Amazon and AllTrails, whose websites/apps I frequently use.

How are Amazon & AllTrails using AI?

On the surface, Amazon and AllTrails appear to be very different businesses – and they are.

Amazon’s users are looking to purchase (mostly) physical products, whereas users of AllTrails seek local outdoor activities, such as hikes and mountain biking routes.

Pictured above: AllTrails (left) and Amazon.com.au (right)

Though they differ on many fronts, the two brands have something in common: they utilise user-generated content (namely, reviews) to help their users make informed decisions.

I recently bought a neck pillow from Amazon for my upcoming Eurotrip (woohoo!) and noticed Amazon have started using AI to summarise customer reviews. Here’s the AI overview that I encountered:

AI in the wild: Amazon now uses AI overviews for product reviews

AllTrails is doing something similar. While deciding on a hike in Tasmania, I stumbled across this:

All Trails is also using AI to summarise reviews of popular hikes

What I love about how Amazon and AllTrails utilisation of AI in their products is their acknowledgement of something crucial: your product is not the destination, it’s a vehicle on a much larger journey.

You want your users to spend as little time with your product as possible and instead enjoy the destination – be that a hike, or a comfortable and stress-free flight to Europe.

Ask yourself, ‘How can I use AI to get my user closer to their destination?’

Some prime examples that I can think of are training LLMs on pre-existing and sprawling knowledge-bases, user manuals (Screaming Frog’s comes to mind), or explaining contracts shrouded in legalese.

In summary, we should look to combine new technology with tried and tested thinking. That is, the user always comes first.

Notes

See Amazon and AllTrails AI overviews in action (I also highly recommend the hike in Tassie; I am yet to try the neck pillow):

https://www.amazon.com.au/cabeau-neck-pillow/s?k=cabeau+neck+pillow

https://www.alltrails.com/trail/australia/tasmania/hartz-peak–4

About the author

Sam Penny is an SEO specialist from Australia. He's interested in the ways that we can use SEO, and search data, to improve the user journey.