The emerging market for autonomous purchasing, often called agentic commerce, depends critically on the ability to determine whether a sale originates from a human shopper or an artificial software agent, and a wave of detection tools is arriving quickly to address this challenge for businesses worldwide.
Proof-of-human systems let a real individual certify the actions of an AI proxy, offering a method to keep malicious bots off websites while avoiding the need for platforms to block all automated traffic, thereby preserving legitimate automation across the user experience and operational efficiency.
World, a startup co-founded by Sam Altman that uses iris scans to confirm human presence, announced this week that its platform can now attach a proof-of-human credential to an AI assistant, enabling the assistant to act on behalf of a verified individual while carrying a verifiable human tag.
“If you know that an agent is coming from a human, you can actually be much more generous,” product chief Tiago Sada said, noting that merchants could soon provide verified agents with special discounts, free trial periods, or priority access to limited-edition items, thereby creating incentives for users to link their identity to their bots.
Online retailers appear poised to accept agent-mediated purchases before other sectors, because the capacity to reward verified bots could reshape pricing strategies, loyalty programs, and overall customer acquisition approaches, potentially lowering friction for automated shoppers while preserving brand control. This shift may also encourage platforms to develop dedicated verification badges and integrate them into checkout flows, creating a new layer of trust for both sellers and consumers.
Meanwhile, several large U.S. banks have begun limiting interactions with third-party agents, citing an inability to confirm which proxies customers have authorized, a development reported by Axios that underscores regulatory pressure to improve agent authentication. Financial institutions fear that unchecked automated access could facilitate fraud or unauthorized transactions, prompting them to demand clearer provenance data and potentially adopt the same proof-of-human mechanisms emerging in commerce.