AI agents aren’t just answering prompts anymore. They’ll soon run treasuries, sign contracts, place trades, and act across chains autonomously. When that goes wrong, the core question is simple: who is accountable for what the agent did? Today, standards like ERC‑8004 help you register agents on-chain and make them discoverable. They tell you an agent exists and what it can do, but not who authorised it or who takes responsibility when it misbehaves Concordium’s Agent Registry was built specifically to close that accountability gap at the protocol level. Every registered agent is cryptographically linked to a Concordium account that belongs to a verified human or entity, via zero‑knowledge identity. This lets agents prove things like if one is allowed to operate in a jurisdiction or one can only spend up to a certain amount, without revealing the underlying personal data. You get real-world accountability and programmable limits, but still preserve privacy on-chain f you believe agents are the new users, you also need a way to know which agents are backed by real users you can hold accountable. That’s exactly the niche Concordium’s Agent Registry is aiming to own in the coming agentic economy. $CCD