FireMon is introducing AI agents to investigate and prepare firewall-policy changes, but they cannot deploy those changes themselves. The security company says customers can limit what each agent can access, and a person must approve any proposal before it is saved to the workflow.
Watch Desk analysis
What happened
SiliconANGLE reports that FireMon Agentic AI will include three agents: one to answer questions about firewall policy, one to recommend rules for an access request, and one to prepare staged policy changes. The last can fill in routine details, but pauses when a practitioner must choose between valid alternatives.
Nothing is saved until a person reviews and accepts the proposal. Even then, it remains in FireMon’s workflow and is not automatically deployed to firewalls. AI features are optional and off by default; customers can choose a supported model, including one hosted in their own approved environment, and restrict each agent to particular device groups.
FireMon also plans a Model Context Protocol server so authorised external applications can retrieve policy evidence and, in permitted workflows, stage selected recommendations. The company says activity is logged for audit. Both the agents and MCP server are due in November, with no additional charge for the initial release.
Why it matters
Firewall access requests can involve several devices and policies, with more than one technically valid way to grant access. FireMon’s pitch is to let agents handle the investigation and paperwork while leaving choices that need organisational judgement to practitioners.
That is a meaningful distinction in security operations: the proposal is not an agent quietly changing live firewall rules. It is an agent preparing work inside defined permissions, with human review before the change enters the workflow. External integrations make the access controls and audit trail especially important, because the system is designed to share policy context beyond its own interface.
Our read
This is a more convincing use of AI than handing an agent the keys and hoping for the best. FireMon describes concrete limits: optional activation, scoped permissions, review before saving and no automatic deployment. The practical test will be whether those controls remain clear when customers connect outside tools and handle complicated requests.
What to watch
- Whether the November release arrives with the stated controls and audit trail.
- How customers scope external applications and agent permissions.
- Whether practitioners find that prepared changes reduce routine work without obscuring the choices that need their judgement.
Discussion spark: For firewall changes, should AI be allowed to prepare and stage proposals for human approval, or should even that work stay in the hands of security practitioners?
Sources and evidence
- Exclusive: FireMon unveils AI agents to handle the legwork behind firewall policy changes – SiliconANGLE (27 October 2026, 12:00 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.