A race is underway among global technology and cryptocurrency companies to build the ideal infrastructure for AI-driven commerce. The surge of advanced protocols over the past three years has rapidly boosted efficiency by standardizing agency-to-agency communication and automating payment processes. These breakthroughs are especially streamlining how software-based agencies share data, trigger payments, and integrate with other digital platforms.
Rise of New Protocols and Market Growth
One standout innovation is Anthropic’s Model Context Protocol, which now records 97 million software development kit downloads per month and operates across tens of thousands of public servers. Google made headlines in April 2025 by handing off its Agent-to-Agent protocol to the Linux Foundation, following swift adoption by more than 100 companies. The Universal Commerce Protocol, unveiled by Google in January, aims to standardize payment and shopping flows between agencies with support from industry leaders such as Shopify, Walmart, Target, Mastercard, Stripe, Visa, and American Express.
On the cryptocurrency front, Coinbase’s x402 protocol has enabled automatic stablecoin payments over HTTP, tallying more than 100 million transactions by the end of 2025. According to a Gartner report, by 2028 one-third of enterprise-grade software applications will feature AI powered by autonomous agents. Deloitte projects that this booming sector could be worth $8.5 billion by 2026, and approach $35 billion by 2030.
The Missing Link: Verification and Arbitration
While current protocols focus largely on establishing agency communication, accessing tools and data, and supporting automated payments, a major challenge remains unresolved. It is still unclear who or what verifies whether a task has been completed correctly and in full. In other words, the critical guarantee layer—ensuring that payment is released only when the work is truly delivered—remains insufficiently addressed.
In February 2026, a proposed Ethereum standard dubbed ERC-8183 outlined a mechanism for job-based transactions in which funds are locked in a smart contract, the service provider submits their completed work, and a third-party evaluator approves or rejects task fulfillment. This approach is designed to deliver programmable collateral and assurance for both human and software agencies.
The Distinction Between Authorization and Verification
Major companies such as Google and Mastercard are prioritizing secure payment authorization, developing protocols reliant on cryptographically signed approval documents. Mastercard’s Verifiable Intent initiative, for example, allows users to securely record and audit purchase approvals. Yet, these systems do not address the need to independently verify whether the promised service or work was successfully completed.
The ERC-8183 draft introduces a requirement for evaluation prior to any payment being made. Evaluators can take many forms—including the client, an oracle network, staking mechanisms, or other decentralized structures. Within the Ethereum community, the ERC-8004 standard has been proposed to serve as a trust and reputation layer for agents and their counterparts.
Security analyses of these protocols warn about the risk of rogue evaluators unfairly approving or rejecting payouts, especially in high-value transactions. As a result, additional safeguards are recommended. In extensive, multi-party networks, the concentration of validation power in a single evaluator has also raised red flags.
Ultimately, as agency-based commerce continues to grow, the distinction between payment authorization and true task verification becomes increasingly apparent. The balance of influence between crypto-backed smart contracts and the authorization infrastructures created by big tech players will be determined as the broader ecosystem develops.
