Digital ID for AI agents: how the NEXTAI launch works
NEXTAI issues verifiable identities to its AI agents: a signed W3C credential, written limits, revocation and a public check that takes 30 seconds.
A digital ID for AI agents is a cryptographically signed credential that identifies a software agent, states which company answers for it, what it is allowed and forbidden to do, when it expires, how it is revoked, and that anyone can check in seconds. NEXTAI, an artificial intelligence consultancy for SMEs operated by Cosmolabs OÜ, has been issuing them since September 2026 for its own agents under the CORTEX Seal brand, with public verification at digitalcortex.tech/verify/argos.
Why now
Recent disclosures about incidents involving agents have put the focus on attribution and permissions. The framework published by OpenAI on 16 September and METR's investigation of 26 August describe cases that call for root-cause analysis and controls. They do not prove any kind of uprising, nor do they let anyone calculate how often the problem occurs inside a given company. The response we propose is to ask for evidence: identity, scope and effective control.
These episodes have different causes and different contexts. The lesson we draw for a company is to ask for evidence of identity, authorisation and operational control; the reports do not allow every incident to be attributed to the absence of a digital ID.
The NEXTAI proposal is simple: before granting an agent access, identify the operator, declare its scope and check how its permissions are applied and withdrawn.
What we have launched
NEXTAI has issued the first verifiable digital identity in its agent fleet to Argos, its commercial prospecting agent, and has published the public identity resources that go with it: a public agentic identity policy, a public agent registry, open technical documentation and a verification page anyone can open without signing up. The full explainer is in Digital ID for AI agents.
As far as we have been able to check, we are not aware of another AI consultancy for SMEs in our market that publishes verifiable credentials for its agents. If one exists, we will be glad to cite it.
| Piece | What it is | Where to check it |
|---|---|---|
| Identifier | The agent's DID: did:web:digitalcortex.tech:agents:argos | DID document |
| Credential | W3C Verifiable Credential 2.0 (vc+jwt), ES256-signed by Cosmolabs OÜ | Credential |
| Signed scope | List of permitted and forbidden actions bound to the manifest hash | Verifier |
| CORTEX Seal | Deterministic image generated from the hash, with a verification QR | Seal |
| Status | W3C Bitstring Status List revocation list; 90-day expiry | Verifier |
| Registry | Every public agent and its status | Registry |
How it works, in six steps
- Manifest. A version of the agent and its SHA-256 hash are registered. Scope and prohibitions are signed as parameters of the credential; they are not to be confused with the effective permissions of each tool.
- Signature. The W3C VC 2.0 credential is secured in vc+jwt format with ES256. The implementation documents key custody in KMS.
- Publication. DID identifiers, the registry and the status of public credentials are published.
- Seal. A visual representation links to the verifier; it can be copied and it does not authenticate the sender.
- Verification. Signature, validity, status and the links of the declaration are all checked.
- Operation. Connectors need permission controls and access withdrawal. That last layer demands evidence beyond the credential.
Autonomy levels: the top level requires human approval
The credential declares identity and scope. NEXTAI uses an internal autonomy scale, which has to be enforced through controls in every tool.
| Level | What the agent can do |
|---|---|
| A0 · Observe | Read, search, gather, detect and measure |
| A1 · Analyse | Classify, summarise, estimate and detect risks |
| A2 · Prepare | Create drafts, proposals, options and plans |
| A3 · Execute internally | Reversible internal changes within its domain |
| A4 · Execute with approval | Persistent or external action only with valid human approval |
| A5 · Prohibited | Not reachable: autonomous money movement, deleting audit trails, privilege escalation |
Argos's credential declares a maximum level of A4 and authorises research, qualification and drafts. It forbids outreach.send, payments, contracts, deletion and the modification of other agents. Approving an action does not lift those prohibitions. The signed scope and the effective permissions of the commercial processes have to be checked separately.
What this is and what it is not
The CORTEX Seal is a verifiable self-declaration about our own agents. The signature protects the declared content, but it does not replace an external assessment, a check of the effective configuration or proof of tool control.
A signed identity does not guarantee quality or security. It declares the operator and the scope; the limits have to be enforced in the connectors. Article 50 of the AI Act sets out transparency obligations according to role and use, but it does not require this cryptographic ID. A timely notice that the user is interacting with AI and the credential itself are complementary measures.
The credential does not prevent an attack or switch an agent off. Containment requires withdrawing access and stopping jobs. Revoking communicates that the identity should no longer be accepted; how long it takes for every consumer to reflect that depends on caches and queries.
What NEXTAI does, and why this fits the rest
For SMEs and mid-sized companies, NEXTAI builds what we call an Enterprise Brain — the corporate memory that AI systems reason over — and Superagents that take charge of complete processes: prospecting, customer service, finance, operations. Agentic identity is not a separate product: it is the governance layer that makes autonomy sustainable.
In practice, every superagent NEXTAI deploys is born with a manifest, limits, an autonomy level, human approval and a credential. The client does not have to take a slide deck on trust: they scan.
Blockchain: what comes next
The current layer does not need blockchain. The roadmap includes studying an anchor on Base with EAS and ERC-8004, with no committed date and no deployment on record. Before announcing it as live, NEXTAI will have to publish contracts, transactions and evidence. An external registry will be able to provide timestamp evidence; it will not prove that a declaration is true or that its limits are respected.
Frequently asked questions
What is a digital ID for AI agents?
It is a cryptographically signed credential that identifies a software agent and declares which company answers for it, what it can do, what it is forbidden to do, when it expires and how it is revoked. It is built on open W3C standards (Verifiable Credentials 2.0, DID, Bitstring Status List) and anyone can check it on a public page without signing up.
Does this mean NEXTAI meets its AI Act obligations?
Article 50 of the AI Act sets transparency obligations according to the role of the provider or deployer and the specific use. It does not require a cryptographic credential and it is not automatically satisfied by a seal. A verifiable identity complements the notices where those apply. See the official text and our guide for companies.
Can the seal be copied?
The image, the credential and its signature can all be copied, because they are public. The signature makes it possible to detect alterations to the signed content. Copying a valid QR does not prove that whoever presents it controls the agent's identity; that requires evidence bound to the interaction.
Does it contain personal data?
No. The credential, the seal and the verifier expose no data about individuals. The responsible party is the company Cosmolabs OÜ, with its Estonian registry code (17106582), an accountable role and a contact channel.
Can another company issue its own identities with this approach?
Yes. It is built on open standards and the specification will be published under an open licence. The technical documentation describes the manifest, the credential, verification and the lifecycle so that any team can reproduce it without going through NEXTAI.
What to do with this
- Open the Argos verification at digitalcortex.tech/verify/argos and look at the eight checks. It is the best way to understand what "verifiable" means.
- Inventory your agents. If you have AI acting on behalf of your company, write down today who answers for each agent, what it is forbidden to do, and who can stop it and in what measured time.
- Ask your provider the six questions. If the answer to "can I check it myself?" is "trust me", that is the problem.
- If you want agents with an ID by design, book your Digital Audit: 45 minutes of technical analysis, no sales decks.
Editorial update of 17 September 2026, produced with the support of AI tools and cross-checked against public documentation. It is not presented as an external audit.
Identity and execution: the scope of this explainer
The public checks describe the credential and its relationship with the registered manifest. They do not prove the authentication of whoever presents it, the state of each external permission, or that actions are actually blocked. See the published evidence and limits.