AI Agents Handle Shopping Autonomously, But Experts Warn of Privacy and Control Risks

What we lose when artificial intelligence does our shopping

Americans spend a remarkable amount of time shopping - more than on education, volunteering or even talking on the phone. But the way they shop is shifting dramatically, as major platforms and retailers are racing to automate commercial decision-making. Artificial intelligence agents can already search for products, recommend options and even complete purchases on a consumer's behalf. Yet many shoppers remain uneasy about handing over control. Although many consumers report using some AI assistance, most currently say they wouldn't want an AI agent to autonomously complete a shopping transaction, according to a recent survey from the consultancy firm Bain & Company. As scholars studying the intersection of law and technology, we have watched AI-assisted commerce expand rapidly. Our research finds that without updated legal measures, this shift toward automated commerce could quietly erode the economic, psychological and social benefits that people receive from shopping on their own terms. Caveat emptor Part of shoppers' hesitation is about privacy. Many are unwilling to share sensitive personal or financial information with AI platforms. But more profoundly, people want to feel in control of their shopping choices. When users can't understand the reasoning behind AI-driven product recommendations, their trust and satisfaction decline. Shoppers are also reluctant to give away their autonomy. In one study involving people booking travel plans, participants deliberately chose trip options that were misaligned with their stated preferences once they were told their choices could be predicted - a way of reasserting independence. Other experiments confirm that the more customers perceive their shopping choices being taken away from them, the more reluctant they are to accept AI purchasing assistance. Although the technology is expected to get better, there have been some well-publicized missteps reported in financial and tech media. The Wall Street Journal wrote about an AI-powered vending machine that lost money and stocked itself with a live fish. The tech publication Wired cataloged design flaws, like an AI agent taking a full 45 seconds to add eggs to a customer's shopping cart. The business case for AI shopping Consumers have good reason to be cautious. AI agents aren't just designed to assist; they're designed to influence. Research shows that these systems can shape preferences, steer choices, increase spending and even reduce the likelihood that consumers return products. And companies are hyping these capabilities. The business platform Salesforce promotes AI agents that can "effortlessly upsell," while payments giant Mastercard reports that its AI assistant, Shopping Muse, generates 15% to 20% higher conversion rates than traditional search - that is, pushing shoppers from browsing to completing a purchase. For companies, the appeal is obvious. From Amazon's Rufus app and Walmart's customer support to AI-enabled grocery carts, companies are rapidly integrating these tools into the shopping experience. Assistants with names like Sparky and Ralph are being promoted as the future of retail, while technologists are calling on companies to prepare their brands for the era of agentic AI shopping. The real concern is not that these systems might fail, but that they may succeed all too well. The human side to shopping AI shopping agents do offer considerable benefits. For example, they can scan numerous products in seconds, compare prices across sellers, track discounts over time, sift through thousands of product reviews, and tailor recommendations to the user's preferences and needs. They can even read through terms of service and privacy policies, helping consumers detect unfavorable fine print. But there's more at stake than these considerations. While consumers have reason to focus on privacy and control, AI shopping agents carry some overlooked emotional risks, such as squashing the joy of anticipation. Psychologists have shown that the period between choosing a purchase and receiving it generates substantial happiness - sometimes more than the product or experience itself. We daydream about the vacation we booked, the outfit we ordered, the meal we planned. Automated buying threatens to drain this anticipatory pleasure. This anticipation connects to another value: a sense of personal and ethical authorship. Even mundane shopping decisions allow people to exercise choice and express judgment. Many consumers deliberately buy fair-trade coffee, cruelty-free cosmetics or environmentally responsible products. The brands and products we choose, from Patagonia and Harley-Davidson to a Taylor Swift tour shirt, help shape who we are. Shopping, moreover, has a communal dimension. We browse stores with friends, chat with salespeople and shop for the people we love. These everyday interactions contribute considerably to our well-being. The same is true of gift-giving. Choosing a gift involves anticipating another person's preferences, investing effort in the search and recognizing that the gesture matters as much as the object itself. When this process is outsourced to an autonomous system, the gift risks becoming a delivery rather than a meaningful gesture of attention and care. Keeping human agency alive AI shopping agents are likely to become part of everyday life, and the regulatory conversation is beginning to catch up, albeit unevenly. Transparency has emerged as a central concern. Past experience with recommendation engines shows that undisclosed conflicts of interest are a real risk. The European Union has proposed a disclosure framework around automated decision-making, although its implementation was recently delayed. In Congress, U.S. lawmakers are considering bills to require companies to reveal how their AI models were trained. So far, consumers seem to want to choose their own level of engagement - a signal that shopping, for many people, is more than just the efficient satisfaction of preferences. Perhaps the least-settled, yet most crucial question is whether AI shopping tools will be designed and regulated to serve users' interests and human flourishing - or optimized, as so many digital tools before them, primarily for corporate profit.

AI "agents" can do your shopping. Should you let them?

Megan Cerullo is a New York-based reporter for CBS MoneyWatch covering small business, workplace, health care, consumer spending and personal finance topics. She regularly appears on CBS News 24/7 to discuss her reporting. Artificial intelligence "agents" promise to do everything from tidying up your email like Marie Kondo declutters a closet, to buying you a pair of heels based on your budget and style preferences. Yet technology experts warn that outsourcing key decisions to AI exposes consumers to risks, potentially leading to communications errors and costing people money, while also potentially handing hackers the keys to their data. This is particularly true when it comes to so-called agentic commerce, or relying on AI agents to make purchases for you. "It isn't mainstream yet and it's pretty risky right now, because there aren't enough guardrails in the system for people to feel comfortable with agents autonomously buying things for them," Matt Kropp, an AI expert with Boston Consulting Group, told CBS News. "It could potentially go buy a car, but I wouldn't say, 'Here's my credit card.'" Such concerns aren't keeping some of America's biggiest companies from charging ahead with AI commerce, which they see as a new way to engage customers and to drive more sales by letting AI do the legwork for shoppers. For example, American Express this week announced new services and protections for cardholders who make purchases using specified AI agents. That includes verifying the identity of an agent when it makes a purchase, according to the credit card issuer. The service "will protect eligible customers from charges related to AI agent error," Amex said in a statement. Amazon's agentic AI assistant, dubbed "Rufus," can track the price of products on the online retailer's platform, alert customers when the price hits a prescribed level and complete the purchase. Walmart, the biggest U.S. retailer, has deployed what it calls a "conversational" AI agent named Sparky that the company says can help consumers find products, provide customer reviews and help with ordering. Roughly a quarter of Americans between the ages of 18 and 39 say they have tried using AI to research products or to shop, according to November data from market research firm Statista. The accelerating adoption of AI is also leading to mishaps. Consider what happened to Sebastian Heyneman, the founder of a San Francisco-based tech startup. According to the New York Times, he instructed an AI agent to secure him a speaking opportunity at the World Economic Forum in Davos, Switzerland. The bot succeeded in landing him a coveted slot at the annual gathering of powerbrokers -- for $30,000, a fee he couldn't afford. Heyneman used a bot by Tasklet, a company that lets businesses automate routine business tasks with AI agents. Andrew Lee, the founder of Tasklet, told CBS News that such problems can arise when a user prompt gives the AI conflicting instructions. Lee also said agentic AI today is fully capable of shopping for people and doing "normal things consumers can do." But just because tech can do something doesn't mean it should be used in that fashion, he warned. "The specific use case of shopping is not a good thing to use these systems for -- yet," he told CBS News. "The agents are fundamentally hard to trust. Personally, I am not super comfortable with that yet. I like to control where my money goes myself, and as a business, we don't recommend that." The reason: Bad actors can lure AI agents into turning over a consumer's personal information, Bretton Auerbach, founder of LocalMovers.com, told CBS News. LocalMovers.com is designed to safely interact with AI agents on behalf of customers, by vetting businesses, booking services, and guaranteeing secure payment to reduce customer risk. "If you give an agent your credit card and say, 'Go to this website and buy me something online,' there are ways to trick the agent," he said. "It might mistake a legitimate website for a phishing website that says in big, bold, text, 'Paste your credit card number here.'"

AI agents won't kill shopping - but agentic commerce could make most of it redundant

And despite what others might tell you, I don't just say that as a mildly-unfashionable middle-aged man who hates everything to do with the experience of buying things. Instead, this question came into focus during a recent discussion with Andrew Bialecki and Chano Fernández, co-CEOs of marketing automation company Klaviyo, about the shift from e-commerce to agentic commerce -- and then further still into what they call "agentic experience." And while their examples are early, they point to a broader shift. Because if agents take over the work of finding, evaluating, and purchasing products on our behalf, then most of what we think of as shopping -- outside of the occasional in-person social experience -- starts to look less like a meaningful activity and more like background infrastructure. No searching. No marketplaces. No defaulting to familiar brands. Which raises an uncomfortable possibility -- what if much of today's online shopping infrastructure exists not because shopping is valuable, but because human cognition doesn't scale to an entire Internet's worth of options? Remove that human processing constraint, and the infrastructure built to scaffold it starts to look less essential -- along with a significant part of today's customer-facing operating models. But what replaces that customer-facing infrastructure? And how might it impact tomorrow's competitive environment? Our story begins with a good old-fashioned Christmas tale -- in this case the growth of agentic commerce over the 2025 holiday period. Many of Klaviyo's customers worked with the company to build agents that could help customers find gifts for their family and friends during one of the busiest periods of the year. Bialecki explains: There's an athletic brand that we work with here in the States, and over the holidays they set up what we call our customer agent. Think of it as like a ChatGPT, but scoped to the business that sits on their website or is available via voice. At first glance, this might simply look like the deployment of an additional channel -- something businesses have already done via mobile. But in practice, the ability to ask questions and add context -- rather than simply browse static information -- doesn't just represent a change in form factor or location, but a change in how customers behave. As Bialecki has observed: When people were browsing the website, they had all these questions like, how does the sizing on this gear run? What kind of climate is it better in, summer or winter? How does it perform when it gets wet? And while much of this information already exists, traditional e-commerce infrastructure places the burden of answering these questions on the customer -- by trawling through product pages, searching for reviews, or browsing forums. But as Bialecki points out, agents can significantly reduce that friction -- especially for less experienced customers: That particular business is targeted more toward non-professional athletes, so they used our customer agent technology to help people discover products, ask questions about them, and ultimately buy -- all through the agent. This means many of the questions being asked are no longer product queries but expressions of need -- a shift in the center of gravity towards the customer's jobs-to-be-done. Practically, Bialecki says he is already seeing this shift in his customer base: We've got a swimwear company in Australia and people aren't just buying swimsuits, but asking questions like: 'I'm going on a trip for ten days, how many swimsuits should I bring?' They're [asking] how to plan a vacation. That's like walking into a shop and getting a helpful shopkeeper who gives you information beyond what's on the shelves. What matters here is not the idea of sales agents per se but the question of what happens to existing go-to-market infrastructure. Because when customers can start with intent instead of products, and expect the system to help them find the right answer, much of the transactional work we think of as shopping -- searching, comparing, narrowing down -- begins to disappear, along with the infrastructure built to support it. And if Bialecki's experience is typical, that trajectory is already accelerating. He tells me: This year we'll go from thousands to probably tens or hundreds of thousands of folks using our agents to deliver customer experience. By the holiday season this year, I think the default experience will be to click to chat with that business. While agents might undermine the basic mechanisms of traditional online shopping infrastructure, they could also do something more fundamental -- remove the need for that infrastructure altogether. And this shift is already starting to show up in practice, as Bialecki describes: Imagine a classic start to the New Year where you go to ChatGPT or your personal agent and say, 'I'm going to train up for a half marathon. How should I do that?' Not only will your agent help to build a plan, but it will help you find a pair of trainers, the right outfit, a water bottle. But what follows from that is the real shift -- agents won't just surface information but will move into the market on the customer's behalf, leading to a shift in how Bialecki sees businesses needing to engage: Rather than looking at websites to figure out what those products are, your agent will query the agents that represent those businesses: 'What's your best trainer for someone starting out? What would you recommend?' Businesses will make their pitch and then the agent will pick the best ones to show to the user. Underpinning this is the emergence of agent-to-agent interaction, where systems act on behalf of both customers and businesses -- with open standards promising to create the conditions for a new form of infrastructure. As Bialecki puts it: There are good open standards, the right building blocks. But what does it mean for a business to offer an agent to every customer? To other agents? There are hundreds of millions of businesses that will need this technology over the next couple of years if they are to be prepared for this agentic evolution. As agents move from channel to infrastructure, it's not only the transactional mechanics of shopping -- browsing, filtering, or comparing -- which have been bypassed. More importantly, agents will hold the most valuable context -- past preferences, current intent, and purchasing authority -- and use that context to make decisions and execute purchases on our behalf. At that point, much of what we think of as shopping begins to look less like a fundamental human activity and more like an inconvenient chore that can be handed off to networks of agents. And the bootstrapping of that agent-to-agent infrastructure is now becoming a cross-industry effort, as Bialecki outlines: We're now working with a lot of the AI labs on what the agent-to-agent experience looks like. Today we often think about agents as something a human interacts with -- but increasingly, we're going to be happy to delegate to agents that pull things together. In this new world, therefore, brands will no longer be competing for human attention within environments they control -- pages, listings, storefronts -- but competing for selection by agents operating beyond their control. Which, as Bialecki outlines, forces a different approach to competition: It's going to be very important that a business gives answers using context about who that person is -- 'I used to run back in university but I haven't done it in ten years. I'm trying to get back into it,' or, 'My friends got me into this. It's my first time.' That could be based on information the customer agent passes through, or information the business already has about its user. You need to be an expert in value-add. Faced with these demands, the familiar structures of e-commerce start to look like artefacts of an earlier model -- a shift Bialecki is already preparing for: We even envision a world where agents are the real source of business truth rather than websites -- with websites being auto-generated on demand based on what you're looking for. Taken seriously, this begins to point to an experience layer that may no longer be anchored in websites, apps, or product catalogs. Instead, experiences become fluid, generated, and contingent on the request being made -- assembled dynamically by a network of agents beyond the control of the participating company. In that world, a significant amount of the work we think of as shopping -- and the infrastructure brands have built to facilitate it -- will be absorbed by networks of agents, forcing brands to compete for selection within systems they do not control. A shift that, if it takes hold, could demand a new technical and operational infrastructure for commercial engagement. What began to crystalize from my conversation with Bialecki and Fernández is this -- most shopping, outside of a small slice of social activity, is a chore rather than a choice. And as a mildly unfashionable middle-aged man, I could have told you that already. In that sense, agents don't kill shopping -- they just reveal that most of it was never worth doing in the first place. But what I find even more interesting is that when we run with the potential implications of agents, it feels like much of the infrastructure we've built to facilitate that unwanted 90% is really just a set of workarounds for a single problem -- the cognitive bandwidth constraints of humans. Brands, search engines, marketplaces, intermediaries -- all of them are, in one way or another, hacks. Mechanisms for capturing, directing, and retaining our scarce human attention by scaffolding our limitations -- reducing the effort required to find what we need through trust signals, discovery shortcuts, or breadth of offering under one roof. But none of that is necessary if agents can simultaneously find and evaluate an infinite number of offerings on our behalf. Instead of needing an aggregation layer to make selection manageable, agents simply become the aggregation layer -- one that is tuned to our goals rather than those of the intermediary. Which helps explain why online giants are now scrambling to own that layer -- and avoid losing visibility into our wants and needs. Because while in the short term, what Bialecki is describing is co-option -- agents being added as just another owned channel for brands to distribute offers and engage customers -- the longer-term implication is potentially far more disruptive. Not a new channel, but a reversal of power. Today, brands compete for attention -- and once they have it, they benefit from the fact that customers don't have the time or context to fully explore alternatives. That constraint creates an asymmetry in which customers rely on brands to reduce uncertainty -- and accept the margin embedded in that convenience because they don't have the capacity to evaluate the whole market. But agents do. With effectively unlimited capacity to search, compare, and evaluate, an agent can see more of the market than the customer ever could -- and more than any individual brand would want them to. At the same time, the agent's evolving understanding of the customer's preferences has the potential to create a new asymmetry -- one which favors the customer over the brand. Because companies are no longer competing for the attention of a constrained human -- they are competing for selection by a system that, if it lives up to its promise, has greater knowledge of the customer's goals and an ability to evaluate the entire market in real time. And that shift -- from optimizing for attention within environments you control to being evaluated within systems you don't -- quietly inverts control away from the brand and towards the customer and their agent. Perhaps not completely. Perhaps not immediately. But directionally. So agents don't destroy shopping as a choice -- for enjoyment, identity, or social signaling. But they do threaten to eliminate shopping as a chore. And in doing so, they may start to erode the e-commerce infrastructures, data moats, and operating models that have grown up around it.

Ecommerce Founders Who Ignore This Type of AI Will Lose Their Best Customers -- Here's Why

Some of the biggest retailers -- Walmart, Target, Home Depot, Macy's, Lowe's, even Amazon -- are implementing AI agents in some way. Small and medium-sized businesses need to, too. There's a term making the rounds in ecommerce circles right now that a lot of business owners are nodding along to without fully understanding, and that term is agentic commerce. If you've heard it and moved on, this is your reminder to go back and pay closer attention. Because what's happening right now is genuinely one of the biggest shifts in how people shop online since Amazon figured out one-click checkout. So let's break it down! Think about how ecommerce has worked for the past 20+ years. A customer has a need. They open a browser. They search. They click around. They compare. They abandon their cart. They come back. They finally buy -- or they don't. That whole messy, human-driven journey? AI is starting to do it for them. Agentic commerce is where AI agents act as autonomous shopping assistants, handling the entire purchase journey on behalf of a consumer. A customer tells their AI assistant what they need, the agent searches across product databases, review sites and brand pages, compares specs, reads reviews, checks pricing and availability, and makes a recommendation or just completes the purchase outright. That last part is the one that should make every ecommerce entrepreneur sit up straight. Business owners sometimes hear "the future of ecommerce" and assume they have a few years to figure it out. Not this time. On Jan. 11, Google CEO Sundar Pichai announced the Universal Commerce Protocol at the National Retail Federation conference, signaling that the time of AI-completed purchases has officially arrived. It's not a beta test; that's actually Google putting a stake in the ground. During Cyber Week 2025, 1 in 5 orders involved an AI agent, for approximately $70 billion in GMV, according to Salesforce. IBM research from January 2026 found that 45% of consumers are already using AI for at least part of their buying journey. Personally, I definitely make up a portion of that statistic. Nearly half. Already. Right now. When Walmart, Target and Google start moving in the same direction at the same time, it's usually a sign that everyone else needs to catch up -- and fast. Brands including Walmart, Target, Home Depot and Lowe's are already partnering with tech providers to build agentic AI solutions. Shopify isn't sitting this one out either. Shopify President Harley Finkelstein, speaking at the Upfront Summit just this week, argued that agentic AI will serve as a new entry point for ecommerce merchants, acting as personal shoppers that bring context to shopping in a way that traditional search engines simply can't. Here's the part that matters for smaller brands: Finkelstein specifically said this opportunity isn't just for the big players. It's for the long tail of merchants, too. Here's where business owners need to get honest with themselves. If an AI agent is out there shopping on a customer's behalf, your brand has to be legible to that agent, not just attractive to humans. AI shopping agents struggle with ambiguity. They need to compare and act quickly. If your delivery windows, shipping costs and return terms are unclear or inconsistent, the agent can skip your offer entirely without a human ever seeing it. Your product pages, your data structure, your pricing clarity; all of it now has to speak to a machine, not just a person browsing on their phone. There's also a discoverability shift happening that business owners need to understand. Brands that previously obsessed over search engine optimization (SEO) now need to become experts in Answer Engine Optimization (AEO), as consumers are turning away from standard keyword searches and using more complex, conversational AI-driven queries. And if your product data isn't visible to those AI channels, you simply won't show up in the recommendations that are driving sales. If you need a number to put this in perspective: The agentic AI market in retail and ecommerce is valued at $60.43 billion in 2026 and is projected to reach $218 billion by 2031. AI-driven traffic to retail sites surged 4,700% year-over-year, with AI-driven visitors converting 31% higher during the 2025 holiday season than other traffic sources. Higher converting traffic coming from AI. That's not a trend to wait on! And yet, only 7% of companies have fully scaled AI, while 62% remain stuck in experimentation, according to McKinsey & Co research. That gap can be your window -- if you take action. Here's what I tell people: You don't have to rebuild your entire business overnight. But you do have to start paying attention to how AI agents are going to find, evaluate and buy from you, because that behavior is already happening -- whether you're ready for it or not. Small businesses using AI are seeing measurable results: 84% of high-tech adopters report gains in sales and profits, and competition is pushing 80% of small businesses to accelerate their technology adoption. The brands getting ahead right now are the ones working with partners who understand both the ecommerce landscape and how automation actually gets implemented at the business level, not just the theory of it. Some Managed Amazon Store companies are already helping ecommerce brands navigate this shift, bringing tools and strategies that were once reserved for enterprise-level retailers down to a scale that actually works for growing businesses. You don't need a massive tech team to take your first steps. But I recommend you clean up your product data. Tighten your shipping and returns language. Think about how a customer would describe what you sell in a conversational sentence to an AI assistant and make sure your store actually answers that. Reverse engineer. Consumer readiness is here, LLM capabilities are maturing and new industry standards are being built to let retailers, platforms and agents work together. The infrastructure is forming around you in real time. The question isn't whether agentic commerce is going to change your business. It already is. The question is whether you're going to shape how it finds you and your business, or just hope for the best.