Building an agentic enterprise means rethinking how you do business. As artificial intelligence (AI) agents start making decisions, enterprises will have to determine not just what agents can do, but also which processes must change to allow agents to function properly.
In payments, that means companies must rethink which parts of their payment model breaks down because human intervention is required.
Traditional buy now, pay later (BNPL) is one of those areas. Over the past 10 years, BNPL has helped retailers solve the affordability problem for countless customers. And by most measures it raises conversion and average order value.
Nevertheless, BNPL was built around a moment that stops AI agents cold, and the companies still waiting on that moment are going to find themselves waiting alone.
In a BNPL transaction, a consumer applies for short-term credit at the point of purchase, receives a real-time approval decision and completes the transaction through a provider like Affirm. This model was created for a human shopper waiting in front of a screen. If there’s a problem, he or she can simply fill out a form, accept the terms and be redirected to a third-party flow.
But now, in the emerging world of agentic commerce, AI agents act on a shopper’s behalf to find, evaluate and complete a purchase. There is no human in the loop to handle a new credit application. The transaction either completes or it doesn’t. With new-credit BNPL approval rates running between 35% and 40%, 6 in 10 BNPL transactions handled by AI could fail at the payment step — and the worst part is, the retailer won’t even know. The agent will just move on to another merchant.
Businesses deploying agents might only discover this limitation in production when the cost of rearchitecting a payment integration is significantly higher than the cost of modeling it correctly from the start.
One emerging solution — installment payments linked to a consumer’s existing credit card — provides the certainty that BNPL lacks. Here’s how it works: the merchant authorizes the full purchase amount against the shopper’s existing credit card at checkout, then splits that authorized amount into monthly payments. That means no new credit application or third-party loan, just installments drawn against credit the shopper already has.
Card-linked installments help prevent payment failures because the credit decision is made when the card is issued. The agent is drawing on available credit, not triggering a real-time lending decision.
Just as important, card-linked models select consumers with established credit and available capacity. These consumers typically make higher-AOV purchases and are repeat buyers. Card-linked BNPL is a high-quality, low-risk growth engine.
In the world of agentic commerce, checkout is where the hard work begins. Retailers who get it right will be the ones agents keep coming back to.
