Arriving fashionably late to the generative AI party, Amazon is planning to incorporate a ChatGPT-style conversational search function into its online marketplace.
With other internet behemoths like Google and Bing already offering (occasionally unhinged) versions of conversational search, and marketplaces like Walmart, Mercari, and Alibaba jumping on the trend, Amazon will likely need to roll out the new feature quickly if it wants to keep up.
While the mega marketplace has not revealed whether it will be overhauling its existing search function, Amazon spokesperson Keri Bertolino told Time that the company would be “significantly investing in generative AI across all of our businesses.”
Conversational search works very differently compared to the conventional keyword-based algorithms you’re used to. But will you have to throw out everything you know about Amazon SEO and start over?
Today we’ll explore what a ChatGPT-style conversational search could look like, and what it could mean for your growing Amazon business.
Amazon’s Future GPT-Style Search
- The State of GPT in E-commerce
- What Does This Mean for Amazon Search?
- How Can Sellers Capitalize on Amazon’s New Search?
- 4 Ways to Prepare for Generative AI Search on Amazon
The State of GPT in E-commerce
You’ve likely heard the buzz about how GPT and related technologies have the potential to transform the e-commerce landscape. Before we dive into what GPT-style search might mean for Amazon sellers, let’s look at some of the basics of AI and e-commerce.
What Exactly Is “GPT”?
GPT stands for Generative Pre-Trained Transformer, a type of natural language processing model that can create text, images, music, and more. It can also interact with humans conversationally, incorporating user responses and feedback into future results.
One extremely popular implementation of this technology is OpenAI’s ChatGPT, a mostly free, text-only application of GPT that any user with an internet connection can converse with. E-commerce businesses can use ChatGPT’s API to create live chat functions on their websites, generate text for product listings, and more.
How Is GPT Used in E-Commerce?
Walmart is already a major player in the generative AI space, using the technology to power several offerings including Text to Shop, and experimenting with GPT for other conversational experiences. Levi Strauss is using AI to create virtual, size-inclusive models for its clothing. And Coca-Cola may be using OpenAI’s technology to optimize its sales and marketing strategies.
But perhaps the most compelling use case for GPT in e-commerce is for product searches and recommendations. That’s because GPT uses natural language processing to incorporate multiple shopper questions and parameters into a single query, while also keeping past conversations, context, and even a user’s data in mind.
Instead of typing a handful of keywords into a search box, consumers can have a text conversation with a chatbot that actually listens.
Generative AI is so disruptive to the way current search algorithms work, it may be poised to overturn e-commerce as we know it. Bill Gates even went as far as to say that, with conversational search, “You’ll never go to a search site again. You’ll never go to Amazon.”
How Is GPT-Style Search Different?
With a standard keyword search, a user has to conduct multiple unconnected searches. Let’s say the user is looking online for a new place to live and work remotely. Each search will be its own little word salad sampler, like “cool weather Latin American city,” “walkable city Andes,” “cost of living South America”, and so on. For each query, the user has to sift through the results to find the best answer.
But with a ChatGPT-style search, a user can type something like: “Where should I move in Latin America if I want a cool climate (60-70 degree days), fast and reliable internet, lots of nature activities, a walkable city with a decent cost of living, good arts scene, and plenty of salsa-dancing options in the evenings?”
Any digitally nomadic sellers reading this will undoubtedly anticipate ChatGPT’s first-choice answer: Medellín, Colombia.
The bot spits out the name in seconds, followed by some basic information about the city, and sometimes a handful of backup options like up-and-coming Cuenca, Ecuador.
How Does GPT-Style Search Work in E-commerce?
While some companies are experimenting with generative AI for more straightforward tasks, like building better product listings, others in the retail space, like Mercari and Instacart, are already implementing early-stage models of AI product search.
With Mercari’s new Merchat AI (currently in beta), for example, a user can type “I need an outfit for salsa dancing in Medellín,” and the bot will respond with a series of queries about size, color, and clothing style, before returning a list of relevant options and photos.
The search results aren’t as impressive as a simple Amazon keyword search for “salsa dancing outfit” — perhaps because Mercari’s algorithm lacks complexity, but more likely because Mercari just doesn’t have the variety of inventory Amazon has at its disposal.
Nevertheless, Mercari’s beta experience provides a good window into what we may see in the future with GPT search.
What Does This Mean for Amazon Search?
Amazon searches operate on the A9 algorithm. The A9 algorithm decides how every product ranks in Amazon’s search results, based on metrics like performance, relevance, availability, price, and fulfillment.
The algorithm is named for A9.com, a former Amazon subsidiary that helped the company develop its search and advertising technology.
What Is the “A10” Algorithm Update?
Technically speaking, there is no such thing as the A10 algorithm — but that doesn’t mean you won’t be hearing about it anyway. “A10” is a widely disseminated misnomer based on the mistaken belief that the “9” in “A9” is a version number.
When e-commerce pundits talk about the “A10 algorithm,” they’re typically referring to a series of updates made between 2020 and 2023, some of which prioritize new criteria in Amazon’s search rankings determinations.
The most recent updates to the algorithm take into account new categories including seller authority, organic sales, offsite traffic, internal sales, and click-through rates.
Amazon and Conversational Search
Amazon reportedly has big plans for generative AI and is likely developing a conversational search feature that could eventually replace its current keyword-driven search box.
According to the language used by the company in (since removed) job posts, the new Amazon Search would include an “interactive conversational experience that helps you find answers to product questions, perform product comparisons, receive personalized product suggestions, and so much more.”
The listings also called the move a “once in a generation transformation for Search,” and compared it to the advent of the Mosaic browser in the 90s.
How Can Sellers Capitalize on Amazon’s New AI Search?
To keep your Amazon rankings on top now and in the future, you’ll want to focus on optimizing for both Amazon’s current search algorithm and a future GPT-style search.
Optimizing for “A10,” a.k.a. the Latest Version of A9
Amazon’s latest search algorithm, which prioritizes a handful of new metrics, could mean big wins in the lead-up to future changes in Search.
Here are the key areas to focus on:
- Seller Authority: Your brand’s reputation — including your seller feedback rating, fulfillment time, and customer service performance — now factor in how highly you rank on Amazon Search. A large, diverse catalog of high-quality products can boost your Seller Authority rating. So can handling customer complaints, replying to comments quickly, and securing good reviews. Winning the Buy Box also delivers a big boost.
- Offsite Traffic: The more traffic you drive to your Amazon listings through external links and social media, the higher you’ll rank with Amazon’s current algorithm. Amp up your social media, affiliate links, and email marketing campaigns to send more shoppers to your Amazon listings.
- Click-Through Rates: Focus on improving your product listings, with clean, helpful text, relevant keywords, and high-quality images. Increase your product reviews and provide a quality product to ensure your reviews will be positive. You can also consider running ads or offering coupons.
- Organic and Internal Sales: Audit and optimize your product listings to make sure they’re doing their job. Look at your pricing and product reviews. Consider using natural language and data structuring (more on this below) so you’ll have less work to do when conversational search becomes a reality.
4 Ways to Prepare for Generative AI Search on Amazon
Searching Amazon with ChatGPT-style AI will look wildly different from the way it does today. But the new search won’t just alter the consumer experience, it will also mean big changes for sellers and their listings.
Here are some insights from industry experts on how you can start to prepare.
1. Write Like a Human and Write for Humans
GPT-style search won’t just crawl your product pages for a bunch of isolated keywords. It will rely on natural language processing to “read” your listings and find clear, relevant answers for users. For the best chance at appearing in their results, put yourself in your potential customers’ shoes and anticipate what questions they might ask.
Answer as many of those questions as you can with well-written, descriptive sentences in terms a real person would use. Use natural language and avoid overly technical terminology.
Users don’t ask “What is the most highly-rated orthopedic lumbar cushion to support the proper curvature of my spinal column?” They ask “What’s the best back pillow?” Or “How can I make my lower back stop hurting?”
2. Think Creatively
“Sellers will need to start thinking about their products in terms of words that describe the experience of using them, rather than just describing specific features,” says Mac Steer, Owner and Director of international SIM card retailer Simify.
Consider the problem the customer is trying to solve and all the ways they might attack that problem. Then gently guide the customer to the solution.
“For example, if someone searches for ‘back support,’ then there’s no reason why you can’t bid on ‘back support cushion’ or ‘back support pillow’ — things that are not necessarily directly related [to the search query], but are part of an overall strategy to help relieve back pain. This is going to require some creativity on the part of sellers,” said Steer.
3. Read Like a Robot
While the content of your listings needs to be natural, you also need to consider how the AI algorithms process information. As with Amazon’s existing algorithm, you’ll need to understand how to optimize your listings from a technical standpoint so that GPT-style search engines can read them more effectively.
Vaibhav Kakkar of SEO and marketing agency Digital Web Solutions says Amazon and other marketplaces may require sellers to make sure their listings conform to specific data formats.
Product listings might need to follow specific structural guidelines, with featured snippets, shorter answers, FAQ sections, and more. Learn as much as you can about key search tactics like structured data and schema markup to help the bots process your listings. “Ultimately, those who adapt quickly will have the greatest chance at success in this new era,” says Kakkar.
4. Consider Future Updates
From voice search to virtual reality (VR), sellers will also need to think about how customers interact with their content, beyond merely reading it off a screen.
Kakkar believes product listings will need to be “informative and engaging enough to capture potential customers’ attention through chatbots or voice assistants like Alexa or Siri, which could be integrated with ChatGPT-style search.”
Think about how your content sounds when it’s read out loud. Take extra care not to bore or overwhelm customers with text that’s really only suited for skimming.
Elisa Bender, Co-Founder of RevenueGeeks, says the new Amazon search is likely to “prioritize products that support VR. If a seller offers a ‘try-it-on’ feature, search engines will give it a higher ranking.”
Stay a Step Ahead with SellersFi
With generative AI forcing sellers to rethink everything they know about Amazon Search, the key is to put the right tools, systems, and processes in place to make sure you’re ready for the new world of search at a moment’s notice.
Now is the time to prepare for the world of AI-powered e-commerce. For help along the way, sign up for the SellersFi newsletter and get free access to the latest Amazon tips and resources designed to help you scale your business. Or get started today and unlock funding for Amazon growth with a free SellersFi account.
