AI shopping is already mainstream. In a 2026 Exploding Topics survey of 1,009 U.S. consumers, 77.6% said they had used AI to help with shopping or purchasing decisions in the previous six months. But the same survey draws a sharp line: shoppers are happy to let AI research, compare, and summarize. They are far less willing to let it spend.
The signal for brands is clear. AI is becoming part of the discovery layer of commerce, especially for product research and price comparison, while checkout still belongs to humans. If you sell online, the immediate question is not whether an agent will complete the purchase for your customer. It is whether ChatGPT and Gemini are shaping the shortlist before that customer reaches your site.
Because it saves time in the exact parts of shopping that people find tedious. According to the Exploding Topics survey, 68.5% of AI shoppers use it for product research and 55.19% use it to find the best price or deal.
Usage is not occasional anymore. The survey found that 43.21% of consumers use AI for shopping weekly or more, and 68.07% of current users said their usage had increased in the last six months. Nearly 69% said AI had definitely influenced them to buy something they otherwise would not have purchased.
The category mix matters too. The survey found especially strong usage in clothing and technology, and 44.62% of respondents had even used AI for grocery shopping. That is a sign of habit, not novelty. People are bringing AI into routine buying decisions, not only big-ticket comparisons.
Tool choice is also consolidating. Among people who use AI to shop, 77.56% said they use ChatGPT and 58.21% said they use Gemini. For marketers, that means AI shopping behavior is not one channel. It is model-specific behavior, and the way your brand appears in each answer can change the outcome.
Trust breaks the moment AI moves from advice to action. Agentic commerce is when an AI system does more than recommend a product and starts taking steps such as storing payment details or placing an order on your behalf.
That jump still feels too big for most shoppers. In the survey, 42.83% said they were not at all aware of direct in-AI checkout tools and another 23.01% were only vaguely aware. Once the idea was explained, the most common reactions were skeptical at 41.08% and suspicious at 33.1%.
The payment layer is even more sensitive. The survey found that 51.45% of consumers were at least somewhat uncomfortable with AI tools storing their card details. More than a third of consumers aged 18 to 29 said they would be very uncomfortable, which undercuts the easy assumption that younger users will automatically accept this shift.
There is also a motive problem. Only 14.16% of respondents believed AI shopping tools primarily serve consumers right now. More people said the tools serve AI companies themselves, 27.52%, or brands and advertisers, 27.32%. If shoppers think the system is working for someone else, they will not hand it the wallet.
Not much. In the Exploding Topics survey, the most common amount consumers would authorize AI to spend autonomously was $0, and the median cap was $50.
The distribution makes the point even sharper. About 31.21% would not allow any autonomous AI spend at all. Another 17.45% would cap it at $20, and 20.74% would cap it at $50. Even among weekly AI shoppers, 51.84% would still cap autonomous spending at $50 or less.
This is why fully automated checkout remains a fringe behavior. The survey found that only 8.99% of AI shoppers use AI as their only source before purchasing, and just 2.18% complete the entire process in AI from research to checkout. Consumers may like AI guidance, but they still want a final review step before money leaves the account.
A useful example from the survey is concert tickets. If an AI agent is meant to buy them the moment they go on sale, that transaction will often exceed $100. Yet only 11.71% of consumers in the survey would currently be comfortable letting AI handle a purchase above that level. Speed is appealing. Surrendering control is not.
It means the near-term battleground is discovery, not delegation. Shoppers are using AI in the messy middle of the journey, where they compare brands, summarize reviews, sanity-check pricing, and narrow options before clicking through or buying elsewhere.
The survey shows an almost even split between two behaviors: 44.8% start on retail websites and then use AI as a supplement, while 44.03% start with AI and then consult other sources. In both cases, your brand needs to survive cross-checking. A strong product page alone is not enough if AI omits you, misframes you, or recommends a competitor more confidently.
Generative engine optimization, or GEO, is the work of improving how your brand appears in AI-generated answers. In this context, GEO is less about chasing vanity mentions and more about controlling how product facts, pricing, comparison points, and brand sentiment get surfaced when a shopper asks for help.
That creates a new operating model for search and ecommerce teams:
This approach works especially well in categories where AI already helps with routine evaluation, such as clothing, technology, and groceries. It shows clearer limits in higher-trust or higher-consideration categories, such as furniture and jewelry, where consumers may still want more human validation.
Probably yes, but on a slower and messier curve than many product roadmaps imply. The same survey that found a $0 typical autonomous spend limit also found that 55.83% of consumers expect AI to play a bigger role in how they shop over the next five years.
That is the paradox brands need to understand. Consumers can believe AI will become more important while still resisting autonomous spending today. In practice, that suggests growth will come first from better discovery, better recommendations, and smoother handoffs, not from bots quietly completing orders in the background.
The survey also found that 55.83% of consumers think AI features make shopping at least somewhat better overall. So the ceiling for AI commerce may rise. It is just unlikely to rise in a straight line from search assistant to fully trusted buyer.
This dataset makes one thing obvious: AI visibility is no longer an experiment for ecommerce brands. If most shoppers use AI for research and comparison, but still make the final decision themselves, then the brand that wins is the one that shows up accurately at the moment intent gets serious.
That is exactly where BotRank's AI Visibility feature becomes useful. Teams can build reusable shopping-intent prompts, run them across models such as ChatGPT and Gemini, and track how often their brand appears, how competitors are described, and which pages get cited. That matters because a shopper who sees your site after using AI is not arriving cold. They are arriving with a preloaded summary, a shortlist, and a bias. If AI left you out, framed you poorly, or cited the wrong page, the conversion problem started before the click. The practical question for brands is not "Will AI buy for my customer?" It is "What does AI say about us when the customer is close to buying?"
Start with the shopping questions AI is already answering well. According to the survey, those are product research, price comparison, gift ideas, brand decisions, and review summaries. That is where demand exists today, and it is where visibility can move revenue faster than speculative checkout features.
The mistake now is treating AI commerce as a checkout feature. The smarter move is treating it as a discovery and perception channel with a growing influence on who makes the shortlist.
Mostly to research them. In the Exploding Topics survey, product research and price comparison were the leading use cases, while only 2.18% said they complete the entire journey inside AI.
ChatGPT led the survey at 77.56% among AI shopping users, followed by Gemini at 58.21%. That makes both platforms important for brands that want visibility during the buying journey.
The survey points to three issues: low awareness of direct checkout tools, discomfort with storing payment details, and skepticism about whose interests these systems serve. When money is involved, convenience stops being enough.
The clearest metric is the spending cap. The most common amount consumers would let AI spend autonomously was $0, and the median cap was $50.
Optimize for AI-led discovery before autonomous transactions. Right now, brands need to win the moments where shoppers ask AI to compare products, summarize reviews, and narrow options before they buy.
