Since the dawn of retail, merchants’ primary job has been to tempt human shoppers to part with their cash. Now they have a new customer to woo: the bots.
It’s still early days, but we are entering the age of agentic commerce, where autonomous artificial intelligence chatbots will be the ones selecting and buying the goods. It’s a shift that has the potential to rewire digital shopping. But for that to happen, as it did with mobile phones more than a decade ago, consumers will need to embrace the new technology and retailers must find a way to engage that doesn’t eat into their profit margins.
The buzz around AI-assisted shopping spiked in September, when OpenAI unveiled its Instant Checkout tool within ChatGPT. It allows US consumers to ask ChatGPT to find a product or for inspiration. Like a virtual personal shopper, the AI engine will present a selection of items, and if the seller in question works with Instant Checkout, the customer can buy and arrange shipping without ever leaving the chat. Retailers including Walmart Inc., Etsy Inc. and Shopify Inc. have signed up. Alphabet Inc.’s Google and Microsoft Corp. also enable people to buy within their AI tools.
Many of the big players are moving swiftly to embrace the new wave of commerce. Wayfair Inc., which competes with Amazon.com Inc. in furniture and home goods, is among the companies that have joined forces with Google, while Britain’s JD Sports Fashion Plc struck a deal with e-commerce provider Commercetools to enable US customers to purchase directly through AI platforms including Microsoft’s Copilot, Google’s Gemini and ChatGPT.
“You really want to be where the customers are,” Oliver Chen, an analyst at TD Cowen, told me.
To thrive in an AI-powered world, companies must ensure that their sweaters or snacks can be found easily by bots and seamlessly popped into virtual carts. Traditionally, brands have built their businesses on understanding and directly serving consumers, so a shift that requires them to interact through AI agents instead is “pretty existential,” Anita Balchandani, who leads McKinsey & Co.’s European apparel, fashion and luxury practice, told me.
One key adjustment companies must make is understanding how large language models look for information. The tools need very clear, precisely structured data to find what a consumer has in mind. That requires retailers to embrace a new, supercharged version of SEO — or search engine optimisation — the key word-driven strategy they now use to make their sites more visible to Google and other search tools.
For example, a person might ask Chat GPT to find them a face cream that is suitable for people with eczema. To appeal to the bot, a retailer or brand should not simply describe the cream on its website as fragrance free. Instead, the description should explicitly state that the product has no scent, additives or preservatives; it should list the ingredients and specify that the cream is a good fit for someone with eczema.
The more conversational commerce becomes, the more important this so-called generative engine optimisation will be.
AI tools take into account not only a retailer’s descriptions, but also third-party mentions from media sites, influencers and reviews. Balchandani notes that bots will pay less attention to aggregated scores, such as star ratings, and instead analyse the substance of reviews — for instance, whether a coat is described as genuinely warm or waterproof.
Retail fundamentals matter too. If a brand’s product is available across a wide range of retailers, it is more likely to show up in bots’ search. Competitive prices also help. And if items are frequently out of stock, then this is going to be a major hurdle.
AI tools have the potential to increase sales. They can offer personalised recommendations and help shoppers discover new-to-them brands. But they also pose significant risk. ChatGPT charges merchants an undisclosed fee for each completed transaction through Instant Checkout. (Google and Copilot are not currently charging commissions on sales that take place via their AI platforms.) If agentic shopping begins to account for a disproportionate share of retailers’ revenue, these fees could start to cut into digital margins, which are already thinner than those of physical stores. And if AI platforms start requiring retailers to pay to show up in searches or to buy ads within chats — the route that Google looks to be taking — that would be yet another drag on profitability.
More fundamental is the fact that companies will be forced to give up some control, as shoppers’ primary relationship will be with the AI platform rather than the store. Many retailers have built lucrative businesses selling advertising to consumer brands and, if they have marketplaces, to other merchants. Those businesses could be imperilled if customers stay within the AI ecosystem, rather than click through to retailers’ websites.
And who will own the data that is generated? Sky Canaves, principal analyst for retail and e-commerce at eMarketer, told me that while the retailer will receive information about the transaction, the AI platform will know much more about how the consumer got there. Eventually AI companies could charge for that knowledge.
The giants, Walmart and Amazon, are big enough to build their own AI tools. Amazon is retaining control over AI in its ecosystem with Rufus, its shopping assistant. But Walmart is keeping its options open, developing Sparky, its own AI helper, as well as pursuing its partnership with OpenAI and working with Google’s Gemini.
It’s not yet clear which approach, if any, will win out. Remember when social commerce was going to be the future of retail? That didn’t play out exactly as promised. While TikTok Shop has taken off, thanks to its unique combination of product discovery, entertainment and ease of purchase, Meta Platform Inc.’s Instagram has never reached its commerce potential.
But with the landscape changing so rapidly, brands and retailers can’t afford to wait and see. Get on the agentic bandwagon now or risk being left behind.
By Andrea Felsted
