Job postings on Amazon’s website earlier this year show that Amazon has plans to add generative AI to its shopping experience.
“We’re working to improve shopping on Amazon using the conversational capabilities of large language models, and are searching for pioneers who are passionate about technology, innovation, and customer experience, and are ready to make a lasting impact on the industry,” one of the job postings read.
What could this new technology mean for brands who sell their products on Amazon? Let’s dive into the likely outcomes from this effort, and how brands can prepare for a new paradigm where consumers are interacting with a conversational chatbot rather than through the existing search bar.
A retail AI chat experience in the wild: Instacart
Instacart, while a great deal smaller than Amazon in terms of gross merchandise volume, has already launched a conversational AI capability in its search function. This gives us some clues as to what Amazon’s AI capability could look like.
Instacart now gives customers the ability to either do a traditional, product-based search, or ask the AI more generalized questions.
I asked the chatbot, What are some ideas for lunchbox meals?
Answer: Here are some nutritious and delicious ideas for lunchbox meals.>
The AI search results shows prepared wraps and sandwiches, fruit cups, and more. These are suggestions that address the intent of my query quite well.
Clicking “go to regular results” shows instead a single option that is a poor answer to the question: raw chicken kabobs.
Instacart says that its conversational AI is based on OpenAI’s ChatGPT AI model, as well as Instacart’s own AI models. Part of Instacart’s unique knowledge graph is its product catalog data: more than 1 billion shoppable items at more than 80,000 retail partners.
Already, we can see the use-case of conversational AI emerging on Instacart. Amazon has a vastly larger product and content data-set to draw on, and I personally expect Amazon to draw on this strength, building a great discovery and consideration experience for consumers.
Likely scenarios for Amazon's consumer-facing AI capability
This technology is likely to enable users to interact with an AI model to discover products and make buying decisions.
The AI will be able to draw on Amazon’s data set and prioritize its recommendations according to factors like product availability, pricing, customer feedback reviews, delivery speed, category relevance, and much more.
With all that Amazon knows about us - our demographics, pet ownership, type of car we drive, what we have been researching and buying lately - its more than possible that Amazon’s AI could give personalized recommendations, allowing them to find the most relevant products suited to their needs.
Personalizing product recommendations and ideas has the potential to really enhance the customer’s shopping experience. If I’m asking for recommendations for dog food, it would be useful to me for Amazon to show results for older dogs rather than puppies - because based on my previous activity, Amazon is likely to know the age of my dog.
Similar to the Instacart model, Amazon is unlikely to completely replace the search function with chat. At least - right away. It’s likely that a chat feature will be made available alongside the existing search bar.
AI within the Alexa voice assistant
Alexa is ripe for enhancements with this new capability. Voice assisted shopping to date has not been widely adopted. Part of the reason for this is Alexas’s limitations in serving product recommendations - much of the time, it was keyword-driven. But AI models are far better at understanding intent, and translating source information into a more helpful answer.
But if Amazon can integrate a smarter, more conversational engine into Alexa, customers may be more inclined to use Alexa to discover and recommend products. With Alexa’s market share in the voice assistant market, this would be a huge win for Amazon.
How could conversational AI change search results on Amazon?
A consumer could ask a question, “what is the best moisturizer for dry skin?”. Using AI, Amazon would be able to utilize all the information that it has to suggest the best moisturizer based on multiple sources with a price to buy a product there and then.
This could potentially make the research phase of a consumer decision journey more efficient in that the customer could get all the information and product recommendations within seconds.
This has many implications on the current way that brands and advertisers manage their products. No longer will a brand need to ensure that they include keywords such as “moisturizer for dry skin” in their product title to prompt the algorithm to put their product first.
Instead, the product needs to include all applicable information in an Amazon and user friendly way including ingredients and how it works to improve dry skin. We can imagine that the support of external sources that recognise the product to be the best for dry skin would also aid the visibility of a product.
What does this mean for brands?
Products that are good at what they do may win greater visibility over products that are just marketed well. This doesn’t necessarily mean that there isn’t room for other products that may not be the best in their category. For example, using AI a consumer could specify that they want the best product under a certain price or without an ingredient which they have an allergy to which may return other products besides what would be considered the best one.
The big to-do for brands is to invest in product content that can be searched by Amazon’s AI model.
This is not new advice, and my strong recommendation is that brands should be investing in product content right now anyway. Ensuring your product content includes all the possible attributes, dimensions, features and benefits that consumers could be looking for - whether in traditional search or in a future AI chatbot - should be a priority for every brand right now.
One ‘unknown’ is how advertising will be integrated. Amazon is the world’s 3rd largest advertising platform, and its ad business is very lucrative. Over time, more sponsored placements have made their way onto Amazon search result pages and even the product listings themselves. It’s hard to imagine that a chatbot will remain completely free and clear of advertising. But Amazon will need to walk the fine line between creating a great and trustworthy customer experience, and serving up those high-margin ad placements.
If you want to go deeper on this topic and learn how to index your product or service in conversational AI results outside of Amazon, download our white paper to learn how existing models like ChatGPT, Bard, and LLaMA 2 source information for their answers, and how brands can use this knowledge to rank for relevant queries.
Amazon's vision statement is "To be Earth’s most customer-centric company, where customers can find and discover anything they might want to buy online."
The mission of Amazon is "to use our technology, selection, competitive pricing and convenience to provide the most customer-centric online shopping experience possible." AI technology on Amazon's website could help them achieve this goal by providing customers with a more efficient, interesting, and personalized way to shop.
And in so many ways, Amazon was built for this. Reams of data, an existing voice assistant that is widely used, one of the best and biggest employers of tech talent on the planet.
While there’s a lot we don’t yet know, the action step for brands is very clear: Ensure your products are correctly optimized for indexing by the AI models that Amazon is building. Do this by maxing out all available opportunities to add content to your products listings and brand store. Try and maximize product reviews (compliantly!). And underpinning it all - deliver a great product experience to your customers