The payments and retail industries are asking the wrong question about agentic commerce. The question is not whether an artificial intelligence agent can complete a purchase on its own, because many agents already can. It is whether the merchant wants to let one do so in the first place.
Forty-six percent of merchants said pricing and the final price a customer pays is the function they are least willing to surrender to agents, according to “Global Digital Shopping Index: The Agentic Commerce Deep Dive,” a PYMNTS Intelligence and Visa Acceptance Solutions collaboration that surveyed 1,185 merchants, 5,241 consumers and 150 acquirers in the United States, Brazil and the United Arab Emirates earlier this year. Fraud, disputes and liability ranked close behind at 42%. Another 41% drew a line around agents redirecting customers to competing merchants.
The report found merchants are sorting AI commerce use cases according to who captures the economic benefit and who absorbs the risk when an AI agent starts making decisions.
That sets agentic AI apart from earlier waves of commerce automation.
Merchants Draw AI’s Red Line at Economic Control
Consumers have already made AI a commerce channel, whether merchants are ready or not. Forty-eight percent of online shoppers used AI while researching their most recent purchase, even as only 23% of merchants can clearly identify both AI-driven traffic and the purchases that follow. Just 15% have structured product data an agent can read.
That gap has not sent merchants rushing to automate checkout. Personalized offers and loyalty rank among their highest-priority investments, with 31% of merchants planning to invest within a year and 62% within two. Automated search, rules-based reordering and ordering built around a customer’s stated preferences follow close behind.
Those use cases share an important characteristic. Each makes an existing customer more valuable without necessarily weakening merchant control.
Agents Become a Negotiating Layer Against the Seller
We’d love to be your preferred source for news.
Please add us to your preferred sources list so our news, data and interviews show up in your feed. Thanks!
Agentic commerce becomes more complicated once optimization starts working against the seller. The agent is not simply another conversion tool. It can become a new negotiating layer between merchant and customer.
Read the report: Global Digital Shopping Index: The Agentic Commerce Deep Dive
Traditional personalization helped a merchant optimize its own funnel. An autonomous buying agent may optimize the consumer’s economics instead by finding a cheaper seller, stacking incentives, switching payment methods or abandoning the merchant altogether.
That is why near-term investment thins as AI approaches the transaction itself. Thirty-one percent of merchants plan to invest within a year in automated product search and comparison, compared with 26% for completing checkout on a consumer’s behalf. Nearly half put capabilities such as multi-merchant bundling, autonomous checkout or dispute management into the “later or never” category.
Consumers Assign AI Authority One Task at a Time
Consumers are drawing their boundaries in the same place. Fifty-six percent will let agents search and compare products, while only 35% are comfortable giving one access to saved payment credentials. The underlying model is not “AI on” or “AI off.” Consumers assign authority task by task.
The winners in agentic retail may not be the companies that give AI the most autonomy. They may be the ones that build the best system for deciding when autonomy is economically useful, when human approval remains valuable and where the agent simply does not get a vote.
At PYMNTS Intelligence, we work with businesses to uncover insights that fuel intelligent, data-driven discussions on changing customer expectations, a more connected economy and the strategic shifts necessary to achieve outcomes. With rigorous research methodologies and unwavering commitment to objective quality, we offer trusted data to grow your business. As our partner, you’ll have access to our diverse team of PhDs, researchers, data analysts, number crunchers, subject matter veterans and editorial experts.
For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.
Add as Preferred Source
