The Dawn of AI-Powered Shopping: How Agents Are Transforming Retail

The future of shopping is quietly taking shape through AI assistants that can do far more than just recommend products. Imagine telling your digital assistant you need a new winter jacket, and it already knows your style preferences and budget based on your past purchases. It searches dozens of retailers, analyzes reviews, checks for sales, and presents you with ranked options. You select one you like, and the AI offers to wait for a price drop. A week later when there’s a sale, it completes the purchase automatically, applies your loyalty points, selects the fastest free shipping option, and sends you a confirmation. Days later, your perfect jacket arrives at your door.

This vision of “agentic commerce” – where AI systems research, compare, and buy on your behalf – has technology giants, startups, and retailers racing to build the infrastructure that will make it possible. McKinsey projects this market could reach a staggering $1 trillion in the U.S. alone by 2030. Major players are already making significant moves: OpenAI has launched a shopping research experience and partnered with Walmart to enable purchases directly within ChatGPT; Google has introduced agentic checkout options; and Perplexity has formed a partnership with PayPal. The momentum is building – Adobe reported that AI-driven traffic to U.S. retail sites surged 670% year-over-year on last year’s Cyber Monday. Despite this growth, we’re still in the early stages, with ChatGPT referrals to e-commerce apps representing less than 1% of sessions over a recent Thanksgiving weekend.

Industry experts note a significant gap between the ambitious promises and current reality of these tools. “I am shocked at the promises versus reality,” says Emily Pfeiffer, a principal analyst at Forrester. Today’s implementations function more like enhanced search tools rather than truly autonomous shopping assistants. Most AI chatbots can suggest products but still require users to click through to retailer sites to complete purchases. While some companies showcase checkout-within-chat capabilities, many polished demos don’t function reliably in real-world conditions. There’s also ongoing debate about whether AI shopping assistants are solving a genuine consumer problem. For categories like fashion, gifts, and home decor—where discovery is part of the shopping pleasure—many consumers may not want an AI to shortcut the browsing experience. Agentic commerce likely will find its most natural fit with low-consideration, commodity purchases like household staples and replenishment items.

The technical challenges of enabling AI agents to interact with e-commerce sites are substantial. Websites were designed for humans typing keywords into browsers, not for AI agents that need to read pages and place orders autonomously. To bridge this gap, several protocols are emerging: Anthropic’s Model Context Protocol (MCP) standardizes how AI agents share context across platforms; Google’s Agent Payments Protocol (AP2) provides a framework for verifiable purchases; and OpenAI, working with Stripe, has developed the Agentic Commerce Protocol (ACP) for completing transactions within ChatGPT. This proliferation of competing standards creates challenges for retailers, who can’t afford to integrate with every AI platform but don’t want to miss potential sales channels. Seattle-area startup Firmly is addressing this problem by developing software that lets merchants connect to multiple protocols through a single interface. “Each protocol is a burden for the merchant,” says Firmly CEO Kumar Senthil, who envisions online retailers needing “microstores” distributed across the internet. Forward-thinking brands like Brooks Running are already refocusing their sites to make them easier for AI systems to read and understand, emphasizing technical SEO through “the lens of AI.”

Amazon, which revolutionized online retail more than three decades ago, finds itself in a complicated position with the rise of agentic commerce. Capturing roughly 40% of U.S. e-commerce spending and operating a fast-growing $70 billion advertising business, Amazon’s current model depends on humans browsing and clicking. The company recently sued Perplexity to prevent the startup from using its AI browser agent to make purchases on Amazon’s marketplace, citing security concerns and a “significantly degraded shopping experience.” Perplexity CEO Aravind Srinivas called this a “bully tactic” and argued consumers should be free to use whatever AI assistant they prefer. Amazon CEO Andy Jassy has acknowledged that agentic commerce “has a chance to be really good for e-commerce” but also criticized current agents for personalization shortcomings and inaccuracies in pricing and delivery estimates. Meanwhile, Amazon isn’t standing still – its AI shopping assistant, Rufus, now has more than 250 million active customers, and the company claims that shoppers using the assistant are 60% more likely to complete purchases. Amazon is also testing a “Buy For Me” feature that lets customers purchase products from other brands’ sites directly within Amazon’s mobile app.

The true potential of AI shopping assistants extends far beyond simply finding products based on specifications. Microsoft’s Kathleen Mitford envisions agents that recognize the context behind purchases – “Imagine an agent recognizing that the bathing suit you’re buying isn’t just another item, but part of preparing for an upcoming vacation and tailoring recommendations accordingly.” This level of personalization would require consumers to share more personal data such as calendars and budget information, but could enable significantly enhanced shopping experiences. John Larson, an investor in Seattle startup Ambassador, believes “real conversational commerce leveraging agentic AI is absolutely the future.” Lorrin Pascoe, CMO at footwear retailer Vessi, echoes this sentiment: “For us, it’s really realizing that this isn’t a gimmick. It is something that is foundational in changing behaviors.” As Vessi opens its first U.S. physical store in Bellevue, Washington – reversing the typical trajectory of brick-and-mortar retailers going online – it’s a reminder that retail evolution rarely follows a predictable path. The same unpredictability applies to agentic commerce, whose ultimate form and impact remain to be seen, though its transformative potential for how we shop is undeniable.