Buying and selling online via traditional e-commerce is old news. The market is entering the age of A-commerce, where artificial intelligence (“AI”) is increasingly able to shop for consumers. At the end of September, OpenAI launched its “Buy it in ChatGPT” trial in the United States, using AI agents built to interact with consumers to do more of people’s browsing and shopping. The technology is known as “agentic commerce.”
ChatGPT already helps millions of people find what to buy. Now it can help them buy it too,” ChatGPT announced late last month. “We’re introducing Instant Checkout in ChatGPT with Etsy and Shopify, and open-sourcing the Agentic Commerce Protocol that powers it, built with Stripe, so more merchants and developers can integrate agentic checkout.” In short: American shoppers can now ask for shopping suggestions from U.S. Etsy sellers within a ChatGPT chat – then buy a product immediately, without having to navigate away to look at individual shop pages.
Looking ahead, big companies in the tech, retail, and financial services segments are touting the next phase of “autonomous agentic commerce.” Experts predict that the emerging shopping model in which AI agents act on behalf of customers could come into effect to a meaningful extent within the next few years.
What’s possible right now? What’s next?
For most people using AI to help them shop, the AI agent is still mostly just searching and recommending products. It still has to shift the customer to the retailer’s website to complete the checkout. For instance, AI can do most steps to order a pizza – though sometimes slower than consumers doing it themselves – apart from paying at the end. That is where individuals still step in: Buyers still need to sign in and/or enter personal and delivery details and then pay.
And that is part of what agentic AI aims to eliminate. With the “Buy it in ChatGPT” trial now underway in the U.S., the customer never leaves the chat, where the checkout is completed. Shopify has said more than 1 million of its merchants will soon be able to check out within ChatGPT, too. Retail titan Walmart has similar plans.
In May 2025, Google launched “AI mode shopping.” Some features, like using a full body photo of yourself to virtually “try-on” clothes, are still only available for U.S. shoppers, with limited brands. At the time, Google said its next step will be a new “agentic checkout […] in coming months” for products sold in the US. It would give shoppers the option of tracking a product until its price drops to within a set budget – then automatically prompting them to buy it, using Google Pay. That checkout option is yet to launch.
Credit card giants Visa and Mastercard are working on ways to make it easier for AI agents to shop for consumers. And American Express is actively building and partnering with other tech and financial companies to enable agentic commerce.
Both the current and coming forms of agentic commerce have the potential to spread fast worldwide, because they run largely on the same global digital infrastructure powering today’s e-commerce: identity, payments, data and compliance. In a report this month, McKinsey revealed that Agentic AI “promises to radically remake the entire shopping experience,” stating that by 2030, the U.S. B2C retail market, alone, “could see up to $1 trillion in orchestrated revenue from agentic commerce, with global projections reaching as high as $3 trillion to $5 trillion.”
“We’re entering an era where AI agents won’t just assist – they’ll decide,” McKinsey asserts.
What are the risks and benefits?
This presents both benefits and risks for today’s commerce ecosystem, per McKinsey. “The transition from static e-commerce to a more dynamic agentic commerce system unlocks transformative potential, redefining customer experiences and operational agility.” In particular, AI agents can create personalized offers that align with a customer’s budget and aesthetic. Subscription-based businesses may see particular gains as agents handle recurring orders automatically, while dynamic pricing systems can fine-tune prices in real time based on demand and competition. Beyond customer engagement, these systems also optimize back-end functions like logistics and inventory, turning AI agents into a force for holistic operational and value transformation.
But these potential gains are not with risks. There are straightforward risks, such as overspending, as agentic commerce removes many steps of the shopping journey found in e-commerce or physical commerce, leading to fewer abandoned carts and potentially higher spending. There is also the risk of fraud, which means that banks will need to figure out how to spot fraud, process refunds, and manage consent when it is not a person pressing “buy,” but an algorithm doing it on their behalf.
Then there are more novel risks, including those that center on customer trust and relationships, since consumers would need to trust AI systems with their private data and preferences, and ensure they are not misused.
Against that background, regulators will need to consider agentic commerce in their competition, privacy, data, and consumer protection rules.
Changing how consumers shop, companies do business
Even for small business, the way products are made and services are discoverable online will have to changein the age of agentic AI. Instead of having websites built for customers and search engines, businesses will need to build AI accessible online stores. Those will not necessarily look like the websites of today; they will be more akin to data-soaked digital catalogues, filled with everything an AI agent needs to place orders: product specifications, price, stock, ratings, reviews, through to delivery options.
Years of bigger brands buying attention and dominating search results might start to matter a bit less if upstarts are able to build good AI accessible online stores. It could be a quiet but massive shift in how trade works. However, each business’s visibility will depend on how AI systems read and rank sellers. If a business’s data is not formatted for AI, it may disappear from view. That could give larger players an edge and once again make it harder for smaller businesses to compete.
Ultimately, this new wave of commerce raises questions for companies and consumers, alike. How willing are consumers to delegate our shopping to AI agents? And are businesses – from brands, retailers and marketplaces to logistics and commerce services providers and payments players – willing and able to adapt to the new paradigm in order to successfully navigate the challenges of trust, risk, and innovation?
Vibhu Arya is a PhD Student in the UTS Business School at the University of Technology Sydney. (Additions courtesy of TFL.)
Share
