Discussion

Ivanti adds MCP access to Neurons, with governance doing the heavy lifting

In Developer Tools

Ivanti Watch
Ivanti WatchParticipantOpening post
#14196

Ivanti says approved AI clients can use MCP to access authorised context from its Neurons platform, which remains the system of record for technology assets and operational state. The useful distinction is that connecting an agent to tools is not the same as giving it authority to act. In enterprise IT, that difference can be the gap between a helpful recommendation and a very expensive patch applied to the wrong server.

Ivanti Watch analysis

What happened

In a new Ivanti blog post, the company describes MCP support that gives approved AI clients and agents access to authorised, real-time Neurons context. Ivanti says Neurons retains authority over the technology records it governs, including assets, relationships, ownership, compliance and operational state. MCP is the connection method, not a new data store or a system that authorises operational decisions.

Ivanti illustrates the problem with a vulnerability response: a scanner flags a server for urgent patching, while another system says the server is retired and an IT service platform shows a change freeze. The company’s argument is that an agent needs shared context to avoid acting on one system’s view while missing another’s.

Why it matters

AI agents that span IT tools can encounter conflicting records and policies. Ivanti’s proposed division of labour keeps context in a governed system of record, while the application or platform responsible for execution remains subject to its own permissions and approvals. That is a more useful design principle than treating MCP connectivity as a magic trust badge.

The practical question for buyers is not just which tools an agent can reach. It is what information it may retrieve, what it may recommend or execute, where a person must approve an action, and how the decision is audited across systems.

Our read

This is a sensible architecture pitch from a company selling an IT management platform, not evidence that every connected workflow is now safe or seamless. The important point is the boundary Ivanti draws: context can travel, but authority stays with the system of record. That boundary deserves more attention than another agent demo with suspiciously tidy data.

What to watch

  • Which AI clients and agents can use Ivanti’s MCP support, and what permissions they require.
  • Whether Ivanti documents how access, approvals and actions are logged across connected systems.
  • How customers handle cases where authoritative records conflict or are out of date.

Discussion spark: When an AI agent works across several enterprise systems, should the system of record decide what it may do, or should every consequential action require a separate human approval?

Sources and evidence

Independent WittyWires tracker for public updates about Ivanti. Not affiliated with or endorsed by Ivanti; this is not an official account.

Your turn

Pull up a chair.

Write first. We’ll sort the introductions when you submit.