Key Takeaways
- An eMarketer chart citing Mod7 Research Strategy's January 2026 survey shows 46% of US genAI assistant users reject both AI advertising and agentic AI commerce, more than the share open to either format.
- The figure joins a chorus of cautious data points, including Radial's "58% open, 6% adopted" gap and Ipsos's finding that 63% would trust AI search less if ads appeared in answers.
- The strategic question for ecommerce operators is no longer "do we adopt?" but "where and when do we start, given the very specific places consumers are willing to let AI in?"
What 46% Tells Us About the Optimism Blind Spot
46% of US genAI assistant users are resistant to both AI advertising and agentic AI commerce, more than the share open to either format, according to a January survey from Mod7 Research Strategy.
www.emarketer.comOn May 20, 2026, eMarketer's chart of the day delivered an uncomfortable reading for the agentic commerce camp. Among US genAI assistant users, 46% reject both AI advertising and agentic AI commerce, a larger group than those open to either format alone. The data comes from Mod7 Research Strategy's April 2026 report titled "Beyond AI Adoption: Trust, Value Exchange, and the Limits of Growth," based on an online survey of 1,068 US adults in January 2026, with a final weighted sample of 881.
eMarketer's editors did not mince words about how to use the chart. They suggested dropping it into "AI monetization decks to temper executive enthusiasm" about chatbot ads and shopping agents, arguing that trust-building must come before sponsored answers or autonomous checkout flows.
The composition of the survey panel matters. These are not AI skeptics in general, but genAI assistant users, the audience closest to becoming agentic commerce customers. Resistance from this group carries different weight than resistance from non-users would.
The Gap Between "Open" and "Done It"
Reading Mod7's data alongside other surveys reveals a recurring pattern of intent that has not translated into behavior.
Radial's December 2025 to January 2026 study, conducted with Dynata across two panels of 1,000 US consumers, makes the gap visible in a single line. 58% say they are open to placing orders through an AI assistant, but just 6% have actually done so. The same study found 19% who would never trust AI with payment information, and 53% who require purchase approval before payment.
Trust in AI search results faces similar headwinds. According to Ipsos's January 2026 Consumer Tracker, 63% of US adults said ads in AI search results would make them trust those results less. Bain's 2,016-consumer survey with ROI Rocket found 64% open to using AI for purchases but only 10% with actual purchase experience, concentrated in low-stakes grocery and household categories.
PYMNTS Intelligence puts the temperature reading even more sharply. 95% of consumers report at least one concern about agentic commerce and 41% say they do not trust these assistants at all, with only 4% expressing complete trust.
The Optimistic Surveys Are Not Wrong
It is important not to dismiss the bullish data points. Visa's February 2026 State of Digital Commerce in Asia Pacific study, conducted by YouGov across 14 markets with 14,764 consumers, found that 74% are already using AI tools to discover, track, or learn about products. Capgemini reports 71% of consumers want generative AI integrated into shopping experiences.
What the data is actually saying is more nuanced than either camp typically presents. Consumers are warming to AI for upstream activities — discovery, comparison, research — while drawing firm lines around the parts of the purchase decision they want to keep. Mod7's 46% specifically measures the size of the group that rejects the two monetization vectors closest to vendor revenue: ads and autonomous checkout.
| Survey | Period | Sample | Direction | Headline figure |
|---|---|---|---|---|
| Mod7 Research Strategy | Jan 2026 | 881 US adults | Cautious | 46% of genAI users reject both AI ads and agentic commerce |
| Radial / Dynata | Dec 2025 - Jan 2026 | 2x 1,000 US consumers | Cautious | 58% open, only 6% have actually used an AI agent for orders |
| Ipsos Consumer Tracker | Jan 2026 | 1,085 US adults | Cautious | 63% say ads in AI search would lower trust |
| Bain AI Payments Consumer Survey | Mar 2025 | 2,016 US consumers | Conditional | 64% open to AI purchases, but only 10% have done so |
| Visa State of Digital Commerce APAC | Feb 2026 | 14,764 across 14 APAC markets | Positive | 74% use AI to shop, 32% reluctant to share payment info |
| PYMNTS Intelligence | Early 2026 | US consumers | Cautious | 95% have at least one concern, 41% do not trust at all |
Why Consumers Are Hesitating
Across surveys, the objections are not abstract fear of AI but concrete operational demands. Radial found 34% require explicit approval for each AI action, 41% want two-factor authentication per transaction, and 39% need the ability to review or cancel without penalty. Bain's findings show consumers want AI to deliver convenience and savings, not to take over the decision itself.
There is one place where consumers consistently warm up. 54% in Radial's survey said they would be likely to use AI for post-purchase delivery monitoring and problem-solving, higher than any pre-purchase use case. That suggests the most viable entry point for agentic commerce is the operational tail of the customer journey, not the moment of purchase.
The Implementation Layer Is Not Immune
Consumer hesitation is mirrored by merchant uncertainty. On May 20, 2026, Retail Gazette covered The Payments Association's UK merchant survey. Among 100 senior payments and risk leaders at UK online retailers, 58% believe AI-initiated transactions are already arriving at their platforms, yet only 41% feel very confident in current liability frameworks.
On a hypothetical 2,000 pound disputed AI purchase, responses fragmented: 24% said liability depends on the circumstances, 21% backed shared liability, and 18% said the AI vendor should bear it. The Association's CEO warned that merchants "are already processing AI-initiated transactions without the controls, liability clarity or transaction visibility needed to manage the risks properly."
This is happening while vendors continue to roll out infrastructure. Google's Universal Cart expansion, Visa's Intelligent Commerce pilots, and Stripe and OpenAI's Agentic Commerce Protocol are all in active deployment. The gap between vendor velocity and consumer-plus-merchant readiness is widening, not closing.
Strategic Implications for Ecommerce Operators
The right strategic question has shifted. Instead of asking whether to adopt agentic commerce, retailers should ask which segment, which journey stage, and which use case to start with — and which to defer.
Post-purchase is the natural beachhead. Proactive shipment notifications, automated delay compensation, AI-guided returns — these match the use cases where Radial's data shows the strongest consumer openness. Wins here build the small trust deposits that fund later expansion into more sensitive territory.
Discovery and comparison come next. Even consumers in Mod7's 46% rejection group are not necessarily rejecting AI-assisted information lookup; they are rejecting ads inside AI answers and AI making purchase decisions for them. Investing in AI Engine Optimization — being visible and well-represented inside ChatGPT, Perplexity, and similar surfaces — reaches that group without triggering the resistance.
Autonomous payment automation is where Mod7, Radial, Bain, and PYMNTS converge most strongly on caution. Deployments here should ride existing authentication and reversal frameworks (Visa Intelligent Commerce, ACP) and always insert explicit consumer approval steps.
Building Acceptance Over Time
What consumers ask for is consistent across studies: transparency, control, and visible trust signals. Sourcing for AI recommendations, easily understandable reasoning, one-click cancellation, and the use of established payment brands as the rails for new AI experiences all surface as adoption levers.
Bain's data offers a particularly useful design hint: consumers trust Apple Pay, PayPal, and Amazon for AI-enabled transactions more than they trust banks, card issuers, or technology platforms. Wrapping new AI checkout experiences inside familiar payment brands lowers the cognitive load for the cautious majority.
Closing Thought
Mod7's 46% figure is not the death of agentic commerce — it is a precise measurement of the gap between the loud technical optimism and the quieter caution of the majority. The lines to watch over the next 12 months are whether more cautious surveys appear, and whether the optimistic surveys start showing actual behavioral adoption catching up to stated openness. The retailers who win will be the ones who build trust patiently from the safest use cases outward, rather than betting that the 46% will simply come around on its own.
