Fetch.ai launches on-chain AEVS for verifiable agent actions

Fetch.ai on May 12 launched the Agent Execution Verification System (AEVS), an on-chain service that issues tamper-evident cryptographic receipts for every AI agent tool call, creating a public audit trail of agent actions.

AEVS records each task an agent performs on the Fetch.ai network — processing refunds, approving payments and coordinating with other agents — and stores a cryptographic receipt on-chain so anyone can verify that the action occurred.

Fetch.ai says the receipts are tamper-evident, so alterations would be detectable, and the on-chain logs are intended to allow independent verification of reported actions.

Market response to the announcement was immediate. FET, Fetch.ai’s native token, rose 3.15% in the 11 hours after the launch while the broader crypto market declined 0.87% in the same window. FET is used for agent discovery, deployment, training and access to machine learning, and it supports AI-to-AI payments alongside USDC. Fetch.ai demonstrated an AI-to-AI payment involving Visa, USDC and FET in December 2025.

AEVS follows earlier platform releases. Fetch.ai released the ASI:One beta on Nov. 19, 2025 to expand multi-agent coordination and interoperability. In December 2025 the company demonstrated AI-to-AI payments. AEVS adds an on-chain verification layer to record and prove actions agents report having taken.

Fetch.ai’s Agentverse hosts more than 2 million agents. As agent interactions scale into the thousands, manual oversight becomes impractical. On-chain receipts provide an audit option that can be checked by users, counterparties or regulators.

Fetch.ai intends AEVS for use cases such as financial approvals, refunds and coordination tasks that require proof of completion. Stored receipts can be used for dispute resolution, compliance checks and third-party audits.

The platform continues to use FET to facilitate agent discovery and economic activity while providing integration options with stablecoins and payment rails demonstrated in prior tests.