AI agents are reshaping retail: brands now pitch to algorithms, not just people

Tyler Ackerman noticed the problem when his team ran a simple test: ask a major AI chatbot to recommend kitchen cabinets, and RTA Store — his family-owned online cabinetry retailer — didn't show up. He spent $15,000 to fix it. He expects to spend up to $12,000 a month on the same problem before long.

Ackerman's predicament is fast becoming every retailer's predicament. Software built into platforms such as ChatGPT, Google AI Mode, and Amazon's Alexa for Shopping is already changing the way consumers find and purchase products. Traffic to U.S. retail sites from AI sources surged 393 percent in 2026, according to Adobe. Morgan Stanley projects that 'agent-influenced spend' could reach $385 billion, or up to 20 percent of total U.S. e-commerce, by 2030.

If we're not focusing on 'How do we have our products read easily by these agents fast and securely?', then we're left behind.— Tyler Ackerman, CEO, RTA Store

The shift is forcing a fundamental rethink of marketing itself. For decades, brands competed for human attention — targeting emotions, aesthetics, and impulse. Now they must also compete for the attention of algorithms that operate on an entirely different logic.

When you advertise to humans, you target eyeballs, emotions and dopamine, but when you advertise to agents, you advertise to logic, a very mathematical logic.— Aviv Shamny, co-founder and CEO, Limy

The scale of consumer adoption driving this pressure is significant. In a survey of 3,054 U.S. adults conducted in August 2026 by Attentive, 68 percent of respondents reported using a general AI chatbot for at least one shopping task during the preceding three months. Across 16 countries, Accenture surveyed 25,590 people and found that 74 percent would place more trust in a personal AI agent than in their best friend when it comes to making a purchase on their behalf. And 62 percent of shoppers, according to data cited by consumer intelligence platform Suzy founder Matt Britton, have purchased a different product or brand than their usual choice because of an AI chatbot recommendation.

That last figure captures the stakes most sharply: a brand absent from an AI assistant's shortlist doesn't just lose a click — it loses the sale to whichever competitor the algorithm surfaced instead.

More and more, you'll just need to click 'buy' in the AI chat interface – you'll never go to a brand's site. The risk of not being prepared could mean you won't show up.— Jeff Blackman, managing director, Barbarian Group

A new category of startup has emerged to help brands navigate this. Companies like New Generation, Gauge, Limy, Catalog, and Evertune are selling services that range from restructuring product data to running millions of test prompts across AI platforms to track how brands appear in chatbot responses. Working alongside Visa, New Generation assesses how prepared a brand's website is to interact with agents, assigning a score from 1 to 100 based on simulated agent navigation attempts.

Say you're a bank and agents are asking about your credit card interest rates. Then a shopper is like, 'Hey, I want to apply for your top-tier credit card' but that run failed because their agent couldn't access the application form. You just lost a customer.— Jonathan Arena, CEO, New Generation

What AI agents actually respond to differs sharply from what human shoppers respond to. In 2024, researchers from Princeton University and the Indian Institute of Technology introduced the term GEO — generative engine optimization — and determined that AI systems drew more heavily from content containing citations, quotations, and statistics. FAQs and bullet points help too, as does regularly updated content. Agents, in the words of one startup founder, 'read' more than human shoppers and prefer content that is 'factual, straightforward and with lots of statistics.'

Instead of saying 'red dress,' it's 'red dress for a party.' It's more contextual.— Dave Stevens, chief technology officer, Groupe Dynamite Inc.

At Aldo, the Montreal-based shoe chain, the team is shifting from simple product tags like 'brown' and 'leather' toward richer content about how to wear shoes with certain outfits. Aldo has also come to recognize that publishing customer reviews — an element currently absent from its website — may be necessary. Matthieu Houle, Aldo's chief information officer, described AI as being "really thirsty for user-generated content."

Because AI models draw on information from across the internet — not just brand websites — every critical blog post and scathing Reddit thread can shape what a chatbot says about a company. That dynamic is elevating the role of public relations inside marketing teams. At Canadian Tire, Ilana Santone, the retailer's senior vice-president of business and AI transformation, says communications executives who previously had a limited presence in marketing meetings are now regarded as "a critical element of that team."

The fragmentation of the AI agent landscape adds another layer of complexity. A single test of 177 toy brands illustrates the problem: when Amazon's Alexa for Shopping was asked for toy recommendations, nearly every result came from Amazon's own store. Ask ChatGPT the same question and it draws from 24 different retailers. Google's AI Mode, posed the same question, pulls from 37 retail sources. Microsoft's Copilot cites retailers most frequently of all, according to tracking by ad firm Tinuiti. A brand distributed solely through independent retailers might go entirely unnoticed on Alexa for Shopping yet show up regularly on ChatGPT — an outcome determined purely by which sources each assistant consults.

Matt Britton, founder of consumer intelligence platform Suzy, calls this emerging discipline 'share of model' — the AI-era equivalent of share of shelf. Retail strategists who once battled for end-cap placement and prime shelf positioning must now track how frequently and how positively each major AI assistant mentions their brand. Britton points out that the key distinction is visibility and negotiability: shelf space could be seen and bargained for through buyer meetings and slotting fees, whereas share of model is determined by data feeds, training sources, and proprietary ranking logic that remains largely opaque to brands.

Vancouver-based luggage brand Monos feeds test prompts — such as 'show me the best carry-on' or 'I'm moving my daughter into her university dorm, what do I need?' — into chatbots to monitor how the brand is represented in their responses. The results vary widely depending on which AI platform is queried, with the brand appearing in responses between 33 and 50 percent of the time across the phrases it tests, according to the Globe and Mail.

Evertune, headquartered in New York and serving more than 100 clients including advertising agency WPP and Canada Goose, processes roughly 150 million prompts a month across several AI models to measure how chatbots surface its clients' brands. Even so, Ed Chater, Evertune's chief product officer, concedes that a portion of this lies beyond retailers' reach, given that AI agents are absorbing information from third-party sources "at a faster rate" than from brand websites.

Not everyone is convinced that chasing algorithmic optimization is the right response. Toby Coulthard, founder of AI marketing platform Jacquard, cautions that the models are evolving "faster than our ability to try to game them." Akshat Trivedi, an industry analyst with FTI Consulting, sees echoes of both the mobile commerce shift and the search engine optimization era — periods when early movers captured outsized gains before the field leveled out — though he notes that the pace this time is different.

The models are changing faster than our ability to try to game them. Maybe these recommendations are going to be true for the next two months but not after that. You have to take it back down to fundamental marketing principles.— Toby Coulthard, founder, Jacquard

Consumer trust in AI agents is real but conditional. Accenture's survey found that while 74 percent of consumers would trust an AI agent more than a best friend to make a purchase, only 9 percent would let an agent act fully independently — initiating and completing purchases on its own. Payments remain the biggest barrier, with only 12 percent currently open to agents making purchase decisions autonomously. But Accenture notes that one in three consumers say a successful low-risk purchase would increase their comfort, suggesting the number willing to delegate fully could grow quickly.

Accenture describes the challenge facing brands as a double victory that must be earned: first with the consumer's heart, and then with the agent's algorithm. Agents reward what they can verify — clean product data, transparent pricing, reliable availability, successful fulfillment. According to Accenture, 37 percent of behaviorally loyal shoppers say they would allow an AI agent to substitute a preferred brand if it identifies a better option. Weak differentiation and inflated pricing, the firm warns, get exposed instantly.

It won't take very long for most of the traffic on the web to be driven by chat. This is going to be very, very transformative to online commerce.— Gil Luria, technology industry analyst, D.A. Davidson & Co.

The financial upside for brands that get this right is already visible in at least one case: Amazon says shoppers who use its AI assistant Rufus are 60 percent more likely to finish a purchase, and the company expects Rufus to generate more than $10 billion in yearly sales. Britton contends that this single figure ought to change the way every retail and consumer goods marketing team approaches platform-specific AI optimization, given that the financial rewards of being recommended well are now demonstrably real.

Trivedi sees the competitive pressure accelerating further. He predicted that "soon we might see products that launch first on AI or discoverability that is driven by brands named in such a way that AI picks them up more than others." Ackerman at RTA Store says the calculation is straightforward: "As these systems get more intelligent, we're going to have no choice but to spend a bit more."

Why it matters — AI shopping agents are already redirecting purchase decisions at scale — and the brands, retailers, and marketing strategies built for a human-browsing internet may be structurally invisible to the algorithms now sitting between consumers and their next purchase.

Reported by nytimes.com, businesstimes.com.sg, theglobeandmail.com, retailtouchpoints.com, forbes.com, accenture.com, mattbritton.com, bain.com