As AI agents increasingly engage in transactions and operate independently online—a market poised to reach $3 trillion to $5 trillion by 2030—an important challenge arises: ensuring that a legitimate human is behind these activities.
World, the identity project supported by Sam Altman, claims to have the answer.
This Tuesday, the company introduced AgentKit, a toolkit for developers that enables AI agents to carry cryptographic proof of being associated with a unique human via its World ID system. The product is compatible with x402, a protocol developed by Coinbase and Cloudflare, which facilitates “agentic payments” by embedding stablecoin micropayments into internet communications, allowing AI agents to transact autonomously.
“Payments are the ‘how’ of agentic commerce, but identity is the ‘who,’” stated Erik Reppel, head of engineering at Coinbase Developer Platform and founder of x402. “This marks a significant step toward a web where agents are recognized not merely as automated traffic but as valid economic participants.”
As AI agents evolve rapidly to address tedious tasks—ranging from making reservations to scouring e-commerce sites for the best deals—estimates suggest that agentic commerce could grow to $3 trillion to $5 trillion by 2030, with agents responsible for as much as 25% of U.S. e-commerce, according to World.
Coinbase founder Brian Armstrong believes that there will soon be more AI agents than humans engaged in transactions, while Binance founder Changpeng Zhao predicts that agents will execute “one million times more payments than individuals,” primarily using cryptocurrency.
The Identity Challenge
With the expansion of the agentic commerce market, the challenge of identity emerges, which payment systems alone cannot resolve.
“An individual could manage thousands of agents, each paying minimal fees,” remarked DC Builder, a research engineer at the World Foundation. “Proof of Human addresses this concern.”
According to a spokesperson for World, AgentKit solves this issue by linking multiple agents to a single verified human, allowing platforms to impose identity-based limits.
“AgentKit enables developers to associate multiple agents with one verified human,” the spokesperson elaborated. “This allows platforms to enforce limits, such as restricting free trials or booking quantities based on the individual behind the agents.”
Moreover, agentic commerce faces resistance, as most websites categorize automated traffic as potentially harmful and often block bots. This strategy clashes with the growing expectation of legitimate software agents acting in users’ interests.
Enhancing Human Verification
AgentKit, presently in its beta version, relies on Orb-based biometric verification, which has stirred controversy. However, the company intends to broaden the system to incorporate additional forms of identification, including NFC-enabled passports and IDs through “World ID Credentials,” enabling users to verify their attributes without disclosing personal data.
“Beyond the beta phase, we aim to enhance AgentKit alongside the next evolution of the World ID protocol,” the spokesperson noted.
With its real-time human verification meter currently reading 17,912,203, World ranks amongst the largest proof-of-personhood networks globally, signaling its ambition to serve as the identity layer of an internet increasingly populated by both people and the AI agents representing them.
Read more: Visa is ready for AI agents. So is Coinbase. They’re building very different internets
