AI agents reshape shopping with seamless journeys
Salesforce data shows that AI and agents drove 20% of all retail sales and accounted for $262 billion in holiday spend during the 2025 shopping season. Brands are responding by implementing agentic tools to personalize experiences, while others are rethinking the traditional e-commerce model entirely — moving from static websites to conversational storefronts that guide shoppers through the full purchase journey.
A new report surveyed 80 brand, retailer and agency respondents about their adoption of agentic tools. Ninety percent said they are already using AI to improve the shopping experience.
“Most people hear ‘agentic commerce’ and think about the discovery moment: how does my brand show up when someone asks an AI what jacket to buy?” said Juan Pellerano-Rendon, Chief Marketing Officer at Swap. “That’s a real and important question to surface in LLMs, but it’s only the entry point.”
He described what happens after discovery as the truly important part. “A real agent doesn’t just surface a product and hand the consumer off to a static website,” he said. “It understands context, it asks questions, it remembers what you looked for last time. That continuity changes the economics.”
How brands are deploying agentic tools
More than half of respondents (56%) are primarily using an internal agentic AI tool, while 33% work with third-party providers. Eleven percent split their efforts between internal tools and outside vendors.
The majority (71%) said their e-commerce sites are set up to be discoverable by AI agents. Twenty percent are unsure, and 9% do not have an AI-discoverable site at all.
Related: Walmart reports soccer fans’ World Cup buying trends
According to the report, more than 1 in 5 brand and agency professionals reported decreases in upper-funnel search traffic (37%) and lower-funnel search traffic (21%) as a result of AI.
“The window to get ahead of this is narrower than most brands realize,” Pellerano-Rendon said. “Discoverability in an agentic world isn’t just about SEO anymore. It’s about whether your product catalog, your brand context and your customer data are structured in a way that an AI agent can actually work with.”
When it comes to engaging consumers, more than 4 in 10 respondents are using AI-native product discovery (46%), in-chat purchasing (46%) and virtual try-on (41%). Personalized recommendations based on user preferences are used by 39%. Less common applications include voice-to-checkout (31%) and hyper-individualized homepages (30%).
Google and Amazon have introduced similar capabilities.
The data challenge behind agentic commerce
Accurate data is needed for agentic AI tools.
“Seventy percent of shoppers leave a site without buying, and a large portion of that abandonment isn’t price or indecision; it’s friction,” the Swap executive said. “They couldn’t find the piece that matched the occasion, the fit, the aesthetic. When an agent closes that gap, when it understands you’re looking for something specific for a specific moment, conversion follows.”
Related: Walmart Tracks World Cup Fan Shopping Habits
The resulting data is completely different from traditional e-commerce. “An agentic storefront generates conversation: what a customer asked for, how they described what they wanted, what they tried on and rejected before they converted,” Pellerano-Rendon said. “That signal is infinitely richer, and most brands don’t have the infrastructure to capture it, let alone act on it.”
Agentic storefronts and the shift from .com to .ai
A majority of respondents are experimenting with or using an agentic storefront in some form, including 40% who are actively deploying one. Another 23% plan to within the next 12 months. Only 7% have no current plans.
Among those with large investments, 62% named improved brand discovery as a leading goal. At least half are also prioritizing increased conversions (57%), improved UX and personalization (56%) and lower customer acquisition costs (50%). Others aim to create an entirely new revenue channel.
“A .com and a .ai can coexist and serve meaningfully different experiences to different consumer segments,” the chief marketing officer said. “As AI becomes further embedded in daily life, the static website — a grid of products, a search bar, a checkout flow — will feel like the mall directory feels today. Functional, but not what you’d choose if you had a better option.”
Adopters anticipate challenges.
Chief among them are data requirements and measuring ROI. Cost is a concern. More than half (53%) worry about low consumer interest or trust, while brand safety and diminished brand identity (35% each) are seen as less big.
Related: Toy Story 5 inspires brand partnerships
“The data we’re already seeing — 2x conversion rates, 3x time on site, 20% reduction in returns — tells us consumers do prefer the better option when they have access to it,” Pellerano-Rendon said. “The question for brands is whether to make the shift early enough to matter.”
Cost and bandwidth slow adoption for some
Among non-adopters, cost and bandwidth are the main obstacles. Only 8% cited insufficient data quality as a reason for holding back. Forty-two percent said they would need lower costs to adopt agentic AI commerce, while current users said lower costs would speed up their investment.
“One of the most common concerns we hear is that brands don’t know if their data is ready,” he added. “The honest answer is, it doesn’t need to be perfect to start. What matters is that you own it. That ownership is the foundation on which everything else is built.”
Regardless of current adoption status, 88% of current adopters and 58% of non-adopters expect their investment to increase over the next 12 months. Twenty-five percent of adopters plan to largely increase spending.
A consulting firm projects the agentic AI market could reach $300 to $500 billion by 2030. Another consulting firm estimates AI agents may generate as much as $1 trillion in U.S. retail revenue by the same year. Analytics firm Nielsen reports 49% of consumers already receive beauty recommendations from generative AI.
“2026 is the year this goes from early adopter territory to an acknowledged strategic priority,” Pellerano-Rendon said. “You’ll start to see brands that launched agentic storefronts in the last 12 months posting results that are hard to argue with, and that will pull the rest of the market forward.”
