2026 will prove whether AI checkout is here to stay

This will be the first full year that shoppers can actually complete transactions within these engines, without having to pay extra for access. ChatGPT, an initiative of OpenAI, has spent the last few months inking checkout deals with all types of companies, from Target to Etsy to Walmart to Instacart. This fall, Shopify also began letting merchants like Skims and Glossier enable checkout in ChatGPT through a function called Instant Checkout. Perplexity, for its part, started offering AI checkout in 2024, but only for Perplexity Pro users; in 2025, it opened this up to all U.S. users and kicked off a partnership with PayPal.

It’s still early innings for AI-powered checkout. For the most part, customers checking out natively on these platforms can only buy one product at a time. They can’t immediately link their loyalty programs or get loyalty points for purchases. Brands don’t get much first-party insight into how customers are shopping for their products on these engines. And they often don’t know who these shoppers are. When users check out through Perplexity, for example, the charge comes back as being from a Perplexity employee.

Still, brands, analysts and technology founders told Modern Retail they’re excited to see what comes of AI checkout in the next 12 months as platforms and retailers iron out details.

“These AI engines are very, very good at doing web search and discovery for you,” said Adam Brotman, a former Starbucks and J. Crew executive and the co-founder of the applied AI company Forum3. “But the second part of shopping is transacting. They’re not yet at prime time, but they’re going to get there. I think next year is going to be more about the year of integration, with the transaction part of the flow.”

The benefits and drawbacks of AI checkout

The retail industry is starting to embrace agentic shopping, in which online agents independently browse and buy products on a customer’s behalf, using their payment information. However, that is still some ways away from mass adoption. Data indicates that some shoppers have concerns about the privacy or data implications of this. Omnisend, for example, found that only 34% of people are willing to let an AI assistant make a purchase on their behalf.

For many customers, clicking the “buy” button themselves within ChatGPT or Perplexity is still a safer bet. When looking purely at the numbers, enabling checkout within these AI engines makes sense for brands and retailers. ChatGPT has massive reach, with 810 million daily active users as of November, according to Sensor Tower.

But there are drawbacks to ceding more control to these platforms. For example, brands don’t get much data from ChatGPT on how people are finding their products.

“It’s kind of a black box,” said Jamie Norwood, the co-founder of Winx Health, which sells sexual health products. “It’s very hard to track how often you’re showing up, and it’s constantly changing.”

Some brands worry that selling through AI engines means they’ll lose out on information they’d have if shoppers came to their sites or apps directly. Bedding brand Brooklinen, for example, wants to be able to integrate ChatGPT with its loyalty program, which it launched in 2020. “I want to know who’s new and who’s a repeat [customer],” said Rachel Levy, Brooklinen’s COO.

Levy shared that customers often find the brand through ChatGPT, but Brooklinen doesn’t know “the purpose” of their search. “Was it, ‘The internet’s favorite sheets?'” she shared at an October media lunch hosted by fraud prevention platform Forter. “We don’t know, and we need to probably do some consumer research to say, ‘Why did you start your journey here?’ to understand.”

Vince Adams, CFO of the fashion brand Everlane, who was also at the Forter lunch, said his company wants to “understand the person behind the click.” “The brand is the one that we feel can make the most authentic connection with [the customer],” he said. “If it’s going to be largely AI traffic coming in, we need to understand that more, in order to build those connections, do post-purchase, loyalty [and] all the things we as brands need to do.”

It’s not a completely new challenge, sources said. When it comes to AI-enabled checkout, retail brands are now finding themselves in a situation familiar to other industries, like air travel and dining, where more customers have started buying from middleman platforms like DoorDash or Expedia.

These aggregators can be powerful sales drivers, but by working with them, companies “effectively gave away the top of the funnel,” said Mikey Vu, who heads the retail AI practice at Bain & Co. “I think a lot of retailers are worried about the same thing happening to them with AI agents,” he shared. Brotman said he could see a day in which OpenAI starts its own loyalty program for ChatGPT shoppers.

Some retail players, wishing to protect their sales and data, have said they will not work with these AI engines. Amazon, for instance, has explicitly blocked ChatGPT’s web crawlers from accessing its product pages, prices and reviews. But Amazon is one of the few companies that can probably afford to put off a partnership with these AI engines, Vu said. “As you go down in size and get to more specialty retailers and smaller retailers, they’re going to have to allow some element of that [to survive],” he said.

Some brands, though, see an opportunity in AI checkout and are looking forward to a day when it can accommodate more complex requests. “The functionality that’s there today is discovery, but can I save my favorites?” Levy asked. “Can I build a cart where I can complete my transaction? Maybe you want to build the entire bedroom, maybe you want to add furniture, maybe you want to add another merchant. I think once that all comes together, it really could change how customers shop and behave.”

Controlling the narrative

Not every brand is directly selling products through ChatGPT, Perplexity or other AI engines. Some aren’t eligible yet. Others, like Winx Health, are in the process of applying. Norwood submitted her email in a ChatGPT-interest form on Shopify’s website and is waiting to hear back on how to participate. “We want to be there,” she said.

Until then, brands are doing what they can to show up in AI results through generative engine optimization, or GEO. Unlike traditional search engines like Google Search, generative engines like ChatGPT pull and rank information based on hyper-specific inquiries from users. These can be anything from, “Find me a dinner under $40 for four people in my city,” to “What’s the best white T-shirt with a pocket on the front?” Whoever shows up first could have financial consequences for brands as more companies start enabling checkout.

While ChatGPT and Perplexity are still a bit of a black box to brands, what companies do know is that, when compiling results, AI engines typically draw from product descriptions, customer reviews, blogs, news articles and even social media tools like Reddit. For brands like Brooklinen, content creation is “the first piece” toward getting ranked highly, Levy said.

“We already spend a lot of time making sure we have a lot of editorial content out there about our products,” she said. If the company gets an award, the engine is “pulling that,” she said. “It’s not pulling the advertising, because you’re paying somebody else for that advertising.”

Similarly, Norwood said she tweaked Winx Health’s Wikipedia page to better show up in AI results. “I added a lot of citations,” she said. “Any article that has mentioned Winx, I linked to it in our Wikipedia.” Norwood also created an awards section on Winx’s page. “I didn’t have that before, just because I was like, ‘Who’s going on our Wikipedia?'” she said. “But now, for ChatGPT or Gemini, it’s important for them to read that we have different accolades.”

Ultimately, checkout via AI is still nascent, but it’s only going to grow bigger as more brands and retailers start working with platforms like ChatGPT, sources said.

“I’m not saying that e-commerce is going to go away — certainly not overnight, and perhaps not ever — but there’s a new channel on the block that we need to consider [with AI],” said Ali Furman, consumer markets industry leader at PwC U.S.

Forum3’s Brotman added, “I will make a prediction that, in 2026, brands that pay attention to the integration layers [of AI checkout] are going to probably have a leg up.”