XDC Network Adds Identity Checks for AI Agents13 min read2,604 words

XDC Network Adds Identity Checks for AI Agents

XDC Network has chosen Concordium to verify AI agent owners while keeping personal data off-chain. See how the system works and what it means for adoption.

XDC AI agent identity verificationXDC NetworkConcordium identity verificationAI agent paymentsinstitutional blockchain adoption
XDC Network Adds Identity Checks for AI Agents

Who Is Behind the Bot? XDC Adds Identity Checks for AI Agents

An AI agent can do more than answer questions. It can find services, manage a wallet, and make payments. That creates an awkward business problem: a transaction can show what happened, but it often doesn't show who's actually accountable for authorizing it.

XDC Network says it's trying to close that gap on XDC AI, its agent payments marketplace, by naming Concordium its exclusive identity verification provider. The integration lets an agent show a publicly verifiable badge that indicates an authenticated person or business stands behind it - without publishing the owner's personal details on-chain.

That distinction matters. Payment capability and accountability are different layers. A wallet can approve a transaction without proving who is behind the software using it. And an identity badge doesn't automatically answer the other questions businesses still care about, like what the agent is allowed to do, how secure it is, or what legal responsibility attaches to it.

This article breaks down how the XDC-Concordium integration works, what it does - and does not - establish, and why it could matter for institutions considering autonomous agents for financial activity.

What XDC Network and Concordium announced

XDC Network has chosen Concordium as the exclusive identity verification provider for XDC AI. XDC AI launched in July 2026 as a marketplace where autonomous agents can discover services and pay for them in USDC. Agents use non-custodial wallets and the x402 payment standard to carry out those transactions.

This new integration adds an identity layer on top of that setup. After verification, an agent can receive an on-chain badge that a third party can check publicly. The badge indicates that the agent's owner - either an individual or a business - completed verification, and the owner's identity information is not published on-chain as part of the badge.

It's easy to miss why this matters if you focus only on payments. A transaction can prove that funds moved from one address to another. It doesn't, by itself, prove which person or organization authorized the activity, controls the agent, or should be contacted if something goes wrong.

As agents gain more authority, the gap grows. An agent that mostly retrieves information creates one set of risks. An agent that can spend USDC, pay for services, or operate inside a business workflow creates a different one. In those cases, counterparties may want evidence that a real, verified owner stands behind the software.

XDC's stated goal is to provide that evidence without putting the owner's personal identity directly on a public blockchain. Concordium's identity framework keeps personal information off-chain. The announcement says disclosure requires a court order and joint action by independent Privacy Guardians - so it's a specific governance approach, not a blanket claim that the data can never be disclosed.

Concordium already supports agent verification on Ethereum and Solana. XDC calls this its first full-network integration. Agents keep their existing XDC wallets, code, and operating setup; Concordium supplies a separate identity verification layer. The agreement also includes revenue sharing tied to verifications, and sponsored network fees mean developers do not need to hold Concordium's native CCD token to use the service.

For builders, keeping identity verification separate from an agent's existing wallet and code can reduce the need to redesign the whole application. It's still not "free" in practice. Teams have to decide how they'll verify badges, what they'll do when verification fails, and how an owner's verification status maps to their own permission and risk policies.

How agent identity verification works

The flow has several steps, and it only works as intended if relying parties understand what was checked and what the badge is claiming.

1. The owner completes identity checks

The person or business behind an agent completes a passport and liveness check through the Concordium Wallet. Concordium says these checks are run by an independent identity provider. The liveness check is meant to confirm the applicant is a live person during verification, rather than relying only on a submitted document.

This step verifies the owner through Concordium's process. It should not be treated as a full assessment of the agent's software, business purpose, or financial risk. Based on the announcement, identity checks are about identity - not a code audit, sanctions screening, ongoing monitoring, or a review of every transaction the agent later makes.

2. The owner links an XDC wallet

After verification, the owner links an XDC wallet using two signatures. The announcement says these signatures don't move funds. Instead, they establish the connection between the verified owner and the wallet involved in the registration process.

This distinction is important. Signing a message to prove wallet control isn't the same thing as authorizing a payment. Still, developers and users should look closely at the signing prompts and the actual implementation rather than assuming every signature request is harmless. Wallet security is part of the overall risk picture here.

3. The agent is registered and receives a badge

The owner registers the agent so it can receive a publicly verifiable on-chain badge. A third party can then check the badge to confirm that a verified owner stands behind the agent, while the owner's personal information remains off-chain.

For a marketplace, business, or service provider, that's a useful signal - but it comes with limits. The badge tells you the result of an identity verification and registration process. It doesn't automatically tell you what the agent is permitted to do, whether the agent's software is secure, or whether the owner is financially liable for a specific action.

The verification service is available via XDC AI's agent verification page, with implementation details in the developer documentation. If you're building against it, those are the places to confirm the current workflow and technical requirements.

Identity is separate from wallet custody

XDC AI agents retain non-custodial wallets. Concordium's role is identity verification, not custody of agent funds. That matters because users shouldn't assume the identity provider is also the wallet provider - or the party approving transactions.

The design splits functions that are often lumped together in product descriptions:

  • Wallet control: the ability to authorize transactions from an address.
  • Identity verification: evidence that an owner completed a specified identity process.
  • Agent permissions: what actions the software is allowed to take.
  • Transaction oversight: the rules, monitoring, and controls applied to activity.

This integration covers identity verification. The other three still have to be handled through wallet behavior, application logic, and broader business processes.

Why accountability matters for institutional adoption

For a consumer experiment, it's often enough to show that an agent can pay. But businesses and financial institutions usually ask different questions: who deployed it, what authority it has, how that authority can be limited, and who will investigate if it behaves unexpectedly.

The XDC-Concordium integration addresses part of that. It links a publicly visible agent to a verified person or business without exposing the owner's personal data on-chain. That can help when a counterparty needs an accountability signal but doesn't need direct access to someone's identity details.

The big idea is straightforward: payment capability doesn't automatically create accountability. Payment infrastructure can let an agent move value, but it can't, by itself, prove who is standing behind the agent. Likewise, an identity badge doesn't make a payment "safe." It doesn't guarantee the agent's instructions are sensible, that the wallet has appropriate limits, or that the owner will reimburse a loss.

A step toward, not a replacement for, governance

If an institution is evaluating AI agents, it will likely need multiple controls working together. Depending on the use case, that can include:

  • A record of who is responsible for deploying the agent.
  • Limits on transaction size, frequency, and permitted counterparties.
  • Approval requirements for higher-risk actions.
  • Monitoring that can spot unusual activity and pause operations.
  • A process for revoking an agent's access or replacing compromised keys.
  • Records that support internal review, dispute handling, and regulatory obligations.

Identity verification can strengthen the "who's responsible" piece. It doesn't cover the rest. For that reason, the integration is best read as accountability infrastructure - not a complete compliance or risk-management solution.

Privacy also deserves careful attention. Keeping personal information off-chain can reduce exposure compared with publishing sensitive data on a public ledger. The announcement says disclosure under Concordium's framework requires a court order and joint action by independent Privacy Guardians. Businesses should still understand what's collected, how it's protected, who processes it, and what procedures apply in the jurisdictions where they operate.

For organizations under privacy rules like the UK GDPR or EU GDPR, a public badge and a private identity record may trigger different data-protection questions. The fact that information isn't on-chain doesn't automatically determine whether the setup meets legal obligations - it depends on the actual data flows, roles, and applicable law.

The commercial model is part of the design

The partnership includes revenue sharing tied to agent verifications. That gives both networks a reason to support the verification service. The sponsored network fees also mean developers don't have to hold CCD just to cover those fees.

These choices may make onboarding smoother, but they don't remove the need to evaluate ongoing costs and operational dependencies. Teams should still ask how verification fees are calculated, who pays them, what happens if sponsored fees change, and how agents behave if identity services are temporarily unavailable. Those are normal infrastructure questions - but they become more important when identity checks sit on the critical path to accessing a financial service.

What developers and businesses should assess

Don't just ask whether an agent has a badge. Ask whether that badge gives your organization enough evidence for the decision you're about to make.

A service that lets an agent browse public information may not need owner verification. A marketplace that allows agents to spend money, access sensitive services, or interact with business systems may need stronger controls. The right level depends on the value at risk, how reversible actions are, and what could happen if things are misused.

Start with the decision the badge will support

Before integrating verification, spell out what changes when you require a verified badge. For example: will a badge be required before the agent makes its first payment? Will it affect transaction limits? Will it unlock access to a restricted marketplace? If the only outcome is that the interface "shows" a badge, the business case may not hold up.

A practical policy could separate three states:

  • Unverified: the agent can use low-risk features, but not higher-value actions.
  • Verified owner: the owner completed the stated identity process, so the agent can use defined permissions.
  • Suspended or invalid: the badge is missing, expired, or no longer accepted, so sensitive actions are blocked until review.

The exact states and rules depend on how you implement it. The key is to define how your application behaves for each state, rather than treating verification as a one-time checkbox.

Test failure cases before production

In real systems, the "happy path" is usually the easiest part. Test what happens when the owner can't complete verification, the wallet link is wrong, a signature is rejected, the badge can't be checked, or the verification provider is unavailable. Decide whether the system should fail closed, continue with reduced permissions, or retry later.

Also plan ownership and recovery. If the person who verified an agent loses access to a wallet, changes roles, or leaves the company, how does your organization update the association? What proof is required? Can an agent's badge be revoked, and how fast do relying applications learn about it? The announcement doesn't specify all of these operational details, so teams should confirm them in the documentation and with the providers before relying on the system.

Keep identity checks in proportion to the risk

Verification adds friction. Passport and liveness checks take time, require the right devices, and involve handling sensitive information through an identity process. For some agents, that cost is worth it. For others - especially if they can't spend money or touch restricted systems - it may be unnecessary.

In most cases, a tiered approach makes sense. Give agents only the permissions they need. Require stronger verification as the potential impact of their actions increases. This avoids two weak extremes: treating every agent as harmless because it's "just software," or forcing every low-risk use case through an unnecessarily heavy process.

XDC and Concordium are aiming at a real infrastructure gap as agents move from generating recommendations to executing transactions. Their integration gives XDC AI a way to attach a verifiable owner signal to agents while keeping personal information off-chain. Whether that becomes meaningful institutional infrastructure depends on how clearly the badge's limits are defined, how well it fits into transaction controls, and how reliably businesses can operate the full system.

If you're building on XDC AI, start with the verification documentation and decide what permissions a verified agent should actually receive. A badge is useful when it changes a well-defined decision - not simply because it appears on-chain.

Conclusion: verification is one layer of accountability

XDC Network's partnership with Concordium adds an identity layer to XDC AI's agent payment marketplace. It lets an agent point to a verified owner without publishing that owner's personal information on-chain. For businesses considering agents that can spend money, that's a more substantial improvement than payment capability alone.

Still, identity isn't the same as permission, security, or liability. Teams need to set spending limits, monitor behavior, plan for revocation, and decide what a verified badge is actually meant to allow. Start with the level of risk your agent can create, then treat verification as one control inside a broader operating model.

Review the XDC AI agent verification guide before integrating it, and test the failure and recovery paths - not only the successful registration flow.

Frequently Asked Questions

It lets agents on XDC AI receive a publicly verifiable on-chain badge indicating that a person or business behind the agent has completed identity verification. The owner’s personal information remains off-chain under Concordium’s identity framework.
The announcement says personal information is kept off-chain and is not publicly revealed through the badge. Under Concordium’s framework, disclosure requires a court order and joint action by independent Privacy Guardians. This does not mean the information can never be disclosed.
No. The badge indicates that an owner has passed an identity verification process. It does not certify the agent’s code, guarantee its decisions are safe, set transaction limits or replace security and compliance controls.
The agreement provides sponsored network fees, so developers do not need to hold Concordium’s native CCD token to use the verification service, according to the announcement. Teams should check the current documentation for operational and fee details.
The verification service is available at xdcai.tech/agent-verification. Developer instructions are available in the XDC AI documentation at docs.xdcai.tech/wallet/agent-verification.

Related articles