Retailers, travel companies and other merchants may soon be getting orders from artificially intelligent agents shopping on behalf of consumers.

Agentic commerce stands to make a new era of shopping by computer possible — one that lets human shoppers tell AI agents what to purchase on their behalf. Then, cyber bots are unleashed to do a consumer’s bidding in the digital marketplace.

That’s a whole new world for merchants and their e-commerce practices, Worldpay Chief Product Officer Cindy Turner said in an interview last week. In the past, merchant sites have been programmed to shut out such bots because all too often they’ve been associated with automated fraud attempts.

Worldpay, and Turner in particular, have been tasked with trying to assist the processor’s clients in preparing for an anticipated wave of bot shoppers. It won’t be an easy shift, but merchants have huge incentives to embrace new age bot-buying because they stand to increase sales if their online systems can distinguish between ‘good’ agentic bots, and ‘bad’ scammer bots.

Agentic commerce spools up

AI agents are expected to drive about $136 billion in consumer-to-business transaction volume generated this year, according to a May forecast from the financial services consulting firm Edgar Dunn & Co. By 2030, the firm predicts that value could skyrocket to $1.7 trillion.

“What's new here is that the agents are entirely AI-based, and the human is no longer present in the purchasing process,” Turner explained. “Without significant intervention by the various payments ecosystem players that introduces a lot of risk to merchants.”

Turner and others envision the AI agents being particularly helpful for planning travel, given the many logistical decisions in that chore, from transportation to hotels to restaurants. Automated grocery shopping may rise as another early use case, given it’s a relatively safe, straightforward and low-cost test of the technology. 

“Travel is one that people believe is going to be frequent,” Turner said. “Inventory replenishment of staples in your closets, in your cabinets – those are the types of things that are low-ticket, low-risk, easy to put AI on top of.”

While such commerce hasn’t mushroomed yet, Turner said she believes it will scale soon. The challenge is to establish a stable business model that balances merchants’ need to block third-party fraud and limit consumer fraud while responding to consumer demand for a tool that helps with mundane shopping tasks, Turner said. 

Aside from Worldpay, other players in the payments ecosystem are stepping in to usher in the new trend, likely recasting the system to make sure the liability for such bots doesn’t fall solely on merchants. Worldpay said this month it would work with Trulioo to give merchants a means of validating agents, authenticating their card credentials and understanding the consumer’s intent for a purchase.

As the model for agentic commerce evolves, the bank card issuers and networks are also likely to weigh in with their perspective on where the liability for purchases that go awry should lie, Turner predicted.

Indeed, Visa and Mastercard are already injecting their ideas into the mix, announcing plans and partnerships earlier this year for enabling agentic commerce. Digital payments pioneer PayPal Holdings has also joined the race to support AI-supported agentic shopping.

Mastercard Chief Services Product Officer Craig Vosburg predicted at a June investor conference that the AI-propelled agentic commerce “is going to progress quickly.”

How payments players could benefit, suffer

While those payments players are doubtless eager to profit from the emerging trend, they’re also on guard against potential disintermediation of their rails, the consulting firm McKinsey noted in a report this month. That’s because not only will the AI agents be able to guide consumers in shopping, they’ll also be able to direct both consumers and merchants to the most economical payment options.

AI agents “can help passive consumers by automatically directing spending to the best card in real time, triggering new applications to snag a better card deal, and rolling balances to another card before promotional rates expire,” the McKinsey report said.

That could hurt issuer and network income by holding down the interchange fees that merchants pay, or by depressing the interest income derived from their cards. “Card issuers face erosion of interchange and interest income when agents rotate cards or default to [account-to-account] rails,”  the report predicted.

Still, other researchers suggest agentic commerce may not blossom the way some expect. For instance, when Turner was asked about the profile of the ideal consumer candidate for AI-driven agentic shopping, she side-stepped the question and noted the new commerce is in a “nascent” phase.

There are naysayers discounting the prospects for computerized agents. Some 40% of agentic AI projects will go by the wayside because they’re too costly, don’t generate real business value or are overly risky, predicted one analyst at the research firm Gartner.

“Most agentic AI projects right now are early stage experiments or proof of concepts that are mostly driven by hype,” Gartner Senior Director Analyst Anushree Verma said in a June statement.