GMO GlobalSign and VESS Labs Complete Proof of Concept for Trust Infrastructure Verifying Identity and Authority Between AI Agents

In anticipation of an era where AI agents autonomously perform product searches, ordering, and payments, the verification focused on a mechanism to confirm which company a connected AI belongs to and the extent of its transaction authority without having to individually query the partner company's systems.

This proof of concept utilizes Verifiable Credentials (VC). This is a mechanism that uses electronic signatures from an issuer to ensure that the certificate's issuer and content have not been tampered with. By verifying the digital credentials presented by an AI agent on the spot, the goal is to use cryptographic technology to confirm the counterpart's identity, authority, and whether human approval exists, even between companies that have not entered into individual prior agreements.

In the demonstration, buyer-side and seller-side AI agents—independently developed by each company—were connected to verify a sequence of processes ranging from mutual identity and authority verification to human purchase approval, ordering, and simulated payment. As of August 28, 2026, they reported that they successfully completed 281 automated tests, including successful transactions as well as cases simulating fraudulent operations such as the reuse of authentication information, use of revoked certificates, and unauthorized modification of purchase details.