What Is DigiCert AI Trust Manager? Identity, Permissions and a Kill Switch for AI Agents

DigiCert AI Trust Manager is a platform for finding the AI agents an organisation uses, giving them a verifiable identity, limiting what they can access and withdrawing their authority instantly when needed. DigiCert announced general availability on 15 September 2026; the product is part of the DigiCert ONE platform.

Why is a product like this needed?

AI agents no longer just answer questions; they connect to systems, read data and start transactions. They carry authority much like an employee account, yet most organisations do not fully know which agent runs where, who owns it and what it can reach.

On the web, the same problem was solved years ago: an SSL certificate proves that a site really belongs to the organisation it claims to be. AI Trust Manager applies the same idea — cryptographic identity built on public key infrastructure (PKI) — to AI agents.

How does it work?

1. Discovery

The platform first finds the agents, models and MCP servers in the organisation and keeps a continuously updated inventory, whether they were bought or built in-house. Approved assets are kept in a private registry and can optionally be published to a public one.

2. AI Passport: the agent's identity

Each agent receives an AI Passport, a portable, cryptographically signed credential. It verifies who the agent is and ties it to an accountable owner — an organisation and a person.

3. “Visas”: the agent's permissions

Alongside the passport come policy-based “visas”. A visa defines which systems, data and actions the agent may access, and for how long. When the period ends, the permission expires on its own.

4. Kill switch

If an agent attempts an action it is not allowed to perform, the action is blocked and its passport can be revoked automatically. Authority is cut at the source, and the agent can be quarantined before it causes harm. Revocation can target a single agent or a group of agents.

5. Records and audit

Activity across agent sessions is recorded. The platform produces a relationship graph linking agents to systems, models, connections and data classifications, along with records that can be used in an audit.

Trust between organisations

One notable capability: another organisation can verify your agent independently, without the two sharing an identity provider. This resembles a browser verifying the certificate of a site it has never seen before; trust comes from a shared certificate infrastructure.

Models and content

AI Trust Manager is the agent-focused part of the broader “AI Trust” framework DigiCert introduced on 30 April 2026. The other two parts are models and content: for models, the aim is to verify that a model has not been altered between development and deployment (the product page lists this as a “preview”); for content, the aim is to sign and verify digital content using the C2PA standard.

Who is it for?

The product targets organisations that run many AI agents and must govern them and keep them accountable. It is not something a single website or a small business needs. DigiCert has not published pricing.

How it relates to SSL certificates

AI Trust Manager is not an SSL certificate and does not replace one. Your website's identity and the encryption of the connection still come from an SSL/TLS certificate. What the two share is the foundation: a cryptographically verifiable identity issued by a trusted authority.

Note: This article is for information only. AI Trust Manager is not part of the DATASSL product catalogue; DigiCert SSL/TLS certificates are available from DATASSL.

Sources: DigiCert AI Trust Manager product page; DigiCert's announcement of 15 September 2026; SiliconANGLE, 30 April 2026.