Korean business daily 매일경제 (Maeil Business Newspaper) reports that an era of "agent payment" has begun — one in which artificial intelligence agents trade goods, and execute payments and transactions on their own.
That framing marks a real shift in what AI software is asked to do. For most of the past few years, AI assistants have been advisers: they draft, summarize, recommend, and hand the final click back to a person. An agent that pays is something else. It completes the last step itself, moving money without a human pressing the button on each transaction.
The change matters because payment is where software stops being reversible. A bad summary wastes a few minutes; a bad purchase moves funds, creates an obligation, and may be difficult to unwind. Once an agent can transact, ordinary questions about consumer software become financial ones: who authorized the spending, what limits applied, who is liable when an agent buys the wrong thing, and how a merchant can tell an authorized agent apart from a fraudulent script.
It also reshapes commerce itself. If shoppers increasingly delegate purchasing to agents, the audience for product pages, advertising, and search rankings changes — the buyer being persuaded may be a piece of software comparing options at machine speed, not a person browsing.
The two available reports, both from 매일경제, describe the arrival of this era rather than detailing specific companies, transaction volumes, or regulatory responses, so the scale of adoption is not established by these sources.
Why it matters: when AI stops recommending purchases and starts making them, everyday questions about trust, authorization, and accountability move from the realm of software quality into the realm of money.