Agentic commerce is already showing up in merchants’ day-to-day business. Customers use AI assistants to research products, and merchants are seeing traffic they attribute to AI tools, but none that sell through their own site or app can yet consistently trace that traffic to a completed sale.
Nine in 10 merchants (90%) observed at least one AI shopping signal in the past 12 months. Three signals tied as the most common, each reported by 85% of merchants: site or app visits believed to come from an AI assistant or AI-powered search, customers arriving having already researched products through an AI assistant, and customers referring to product information they got from an AI assistant. Only 10% of merchants received requests from AI platforms or agent providers to work together.
Merchants also report that AI-driven product discovery is already benefiting their businesses, although the gains are uneven. Overall, 82% of merchants say AI-driven product discovery has helped their business, and none report a negative impact. The benefits cluster among merchants already growing: 91% of those whose sales rose last year report a positive impact, compared with 50% among those whose sales declined. Merchants with a higher share of online sales also appear more likely to benefit.
Merchants Struggle to Connect AI Traffic to Sales
AI attribution remains a major infrastructure gap.
Merchants are gradually improving their ability to identify AI traffic and trace it through the sales funnel. The picture remains far from complete, however, leaving them to make decisions about AI visibility, investment and partnerships without a clear view of what that activity is delivering. That makes attribution the first major infrastructure gap in preparing for agentic commerce.
No merchant that sells through its own site or app in the sample can fully and consistently connect AI-driven visits to completed purchases. About half (51%) can identify some AI-originated visits, but attribution remains inconsistent. Another 37% know a visit came from an AI platform but lose the trail before purchase, and 12% cannot separate agentic and human traffic at all.
Merchants prioritize AI visibility over attribution.
Merchants are prioritizing visibility over analytics as they ramp up for agentic shopping. Eighty percent call AI visibility a high or critical priority, compared with 62% who say the same about identifying AI-driven visits and purchases. The pattern recalls the early SEO era, when presence often came before reliable attribution. While logical from the perspective of driving sales, this approach risks leaving merchants unable to determine what their investments are actually delivering.
Most merchants are already working on the visibility side. Eighty-seven percent have taken at least one step to improve how their products appear inside AI platforms, and just 5.0% have no plans. The most common moves focus on product content: 85% have improved titles, descriptions, attributes or imagery for AI, and 73% monitor how their brand or products appear in AI-generated answers. Merchants invest an average of roughly 2.0% of revenue in AI visibility and measurement combined, or about $7 million annually for a merchant near the sample midpoint, although more than half still invest less than 1.0%.
Agentic Commerce Is Set to Scale Quickly
Merchants expect agentic commerce to reshape sales.
Merchants expect agentic commerce to significantly affect commerce. Two-thirds (67%) expect AI agents to significantly impact commerce, and another 17% predict a fundamental transformation. Just 2.0% see little or no lasting effect. Most also expect the shift to occur quickly: Three in four believe it will arrive within two years, including 22% who say it already has or will within a year.
Merchants also expect agents to take a meaningful share of sales. On average, merchants predict that AI agents will complete 13% of digital sales within two years, and 70% expect agents to account for at least 5.0%. Expectations rise with readiness. Those that describe themselves as agent-ready expect agents to handle 28% of digital sales in two years, compared with 7.0% among merchants still in the early stages.
Merchants see discovery gains but margin risks.
Merchants expect some of the clearest benefits to come on the demand side. More than four in five expect agentic commerce to help them reach new or niche customer segments and improve product discovery. Sales volume and brand recognition also rank highly, suggesting merchants see agents as a potentially powerful new channel for demand generation.
The picture is much less optimistic—and in some cases points to a downside risk—when merchants consider margins and customer relationships. Only 42% expect a positive effect on profit margins, while 18% expect margins to suffer, the weakest net outlook of any dimension tested. Customer loyalty and repeat purchases also rank near the bottom. Among the 13 merchants expecting any downside from agentic commerce, more than two-thirds point to margins.
Preparing for Agents Without Giving Up Control
Merchants are preparing but may underestimate rivals.
With merchants overwhelmingly expecting agentic commerce to influence sales and traffic, nearly all are taking steps to prepare for AI-led shopping. Nine in 10 have begun readying their catalogs for agents, and 87% have worked to improve how their products appear on AI platforms, with 87% doing both.
At the same time, merchants appear to widely underestimate how quickly their rivals are moving. Only 15% believe that most or all direct competitors are preparing for agentic commerce, while 60% say “some” and 20% “few” or “none.” That perception gap may be giving merchants a false sense of security about how urgently they need to prepare.
AI agents threaten merchants’ control.
While merchants clearly see agent-driven shopping as an opportunity, they are well aware that it presents a threat to their control of traffic and customer relationships. Ninety-three percent expect AI agents to take traffic from at least one channel they use, led by their own website (86%), search traffic to their site (77%) and their own mobile apps (76%). Notably, even 23% of merchants with physical stores believe in-person traffic will be impacted.
To balance the need to increase agent readiness with the perceived risk of giving up too much control, 86% of merchants that sell through their own site or app plan to take a hybrid approach over the next two years. They will either keep direct traffic primary while staying agent-accessible (49%) or weight direct traffic and agent commerce about equally (37%). Few are choosing the extremes (2.0% agentic-first, 5.0% traffic-first), and 7.0% have no defined strategy. The dominant approach is a hedge: Remain accessible to agents without ceding the direct relationship and the customer data that comes with it.
Merchants are limiting what agents can buy.
Another way merchants plan to manage agentic shopping is by limiting which products agents can buy. About six in 10 merchants would hold back high-value or big-ticket items, and roughly half would protect custom or personalized products, where mistaken purchases can be costliest to unwind. Among merchants willing to expose the full catalog, more would do so on different terms than on the same terms as their direct channel. Fewer than one in six have actually defined which products agents may buy.
Buying Into Agentic Commerce
Most merchants do not expect to build the agentic commerce stack themselves. Instead, the market is taking shape around vendors and platforms that can add agent capabilities to systems merchants already use.
Nearly two-thirds plan to buy: 52% would purchase point solutions and integrate them, while 13% would buy a bundled solution. Another quarter would wait for their existing platform provider to add the capability, compared with just 2.0% that plan to build in-house and 8.0% that remain undecided. That leaves the opportunity concentrated among outside providers, especially those already embedded in merchants’ eCommerce operations.
Six in 10 merchants would rely primarily on their eCommerce platform provider for support or guidance in preparing for agentic commerce. That is more than three times the combined share naming a bank, processor or acquirer (15%) or a payment network (3.0%), pointing to a blind spot for the commercial banking and payments space. AI platforms themselves rank near the bottom, at 5.0%.
Merchants also expect participation itself to carry a cost. Eighty percent believe merchants will pay a fee or commission on completed agent sales, and 92% expect some form of payment once subscriptions and other models are included. Taken together with the overwhelming share already preparing for agentic commerce, this suggests merchants are willing to pay to participate—and raises the stakes around which providers become their default partners.
Read More
PYMNTS Intelligence is the leading provider of information on the trends driving innovation in finance, banking and digital payments. To stay up to date, subscribe to our newsletters and read our in-depth reports.
Methodology
“The Agent Advantage: How Merchants Plan to Win AI Shopping” is the first edition of the PYMNTS Intelligence Merchant AI Report series. This debut edition is based on a survey of 60 senior eCommerce decision-makers at established, digitally active U.S. merchants that sell to consumers and have annual revenue between $50 million and $1 billion. The survey was fielded in August 2026. Given the focused sample size, findings are best read as directional signals from the merchant segment closest to the agentic commerce frontier. The report examines how merchants understand agentic commerce, the AI shopping behavior they already see, the impact they expect on their business and the strategies they are deploying to prepare.
