What we still need to verify : 3 points in this profile are not yet confirmed against vendor documentation.
- Scope and maturity of the AI agent identity capabilities specifically: verify against vendor docs
- Supported industrial protocols and device classes: confirm
- Integrations with enterprise identity providers and secrets platforms: confirm
Treat these points as unconfirmed. They are open items in the catalog's verification queue, and this note stays until each is checked against the vendor's documentation.
What it does
Xage is a zero trust access platform built for environments the usual enterprise identity stack does not reach. Its origin is operational technology: industrial control systems, field assets and legacy equipment that cannot run an agent, cannot speak modern authentication protocols, and assume anything already on the network is trusted. Xage places enforcement in front of those assets and brokers every interaction, so identity, authorization and multi factor challenge apply even to devices with no concept of a user.
The distinguishing architectural choice is decentralization. Policy and identity state replicate across a mesh of nodes rather than sitting in one directory, backed by a tamper evident distributed ledger so no single node is the point of failure or of compromise. That matters where sites run intermittently connected and a central broker outage cannot be allowed to stop a plant. The platform also rotates credentials for accounts that would otherwise be shared and static, and brokers remote access for vendors without exposing the network.
The AI relevance is extending the same model to non human identities. An agent acting on systems is an identity that must be enrolled, scoped, challenged and revoked like a contractor, and the enforcement point already mediating asset access is a reasonable place to do that. Treat this as identity infrastructure applied to agents rather than an AI specific product, and verify how far that capability has gone.
Where it fits
At the access boundary in front of assets, operated by the OT security or infrastructure team with site operations. It is infrastructure, not an add on: adoption means placing enforcement nodes into the environment and moving access paths through them. That requires an accurate asset inventory and a map of who and what connects to each system, which many industrial environments lack.
Strengths
- Applies identity and authorization to equipment that structurally cannot authenticate on its own.
- Distributed enforcement survives site isolation and network partition, which centralized brokers do not.
- Replaces shared static credentials on industrial assets with managed, rotated ones, closing a persistent exposure.
- Secure remote access for third party engineers removes a poorly controlled ingress path.
Limitations
- A deployment project, not a tool you switch on. Expect asset discovery, network design and change management before value appears.
- Inserting an enforcement point into a control system path carries availability risk that plant operations will scrutinize hard, and rightly so.
- The agent identity story extends an existing platform, so confirm its depth. Little of this applies to teams securing software they build.
Who it suits
Operators of industrial, energy, utility or manufacturing environments needing enforceable access control over assets that predate modern identity, and able to run an infrastructure program. Not appropriate for a software organization looking for guardrails around AI features in its products.
Used Xage Security? Recommend it under your own name and title.
Recommend this tool