See how the agent detects the root cause of a VPN outage impacting 47 users, deploys the fix autonomously and updates the knowledge base — in 3 seconds.
The IT Service & Operations Agent is a Claude agent that handles level 1 and 2 support tickets: triage, root-cause search, applying the documented fix and updating the knowledge base. It escalates complex or sensitive incidents to your teams with full context, and every action on your systems is logged in LOOP™.
The agent works for IT support and operations teams. It absorbs the level 1 and 2 requests that come back every day, so engineers can focus on complex incidents and projects. It connects to your ITSM, monitoring and knowledge base through the MCP protocol, under your security policies.
Tickets from your ITSM tool, monitoring alerts, the directory of users and devices, and your knowledge base with its resolution procedures. It also knows the scope the IT department has given it: the systems it may act on and those it must leave to the teams.
It qualifies each ticket, groups requests that belong to the same incident and looks for the cause using alerts and history. If the fix is documented and within its scope, it applies it and checks that the service is back. Otherwise, it routes the ticket to the right team with its diagnosis and what it has already checked.
Resolved, documented tickets, users kept informed of progress, a knowledge base that grows with each new resolution and, for escalated incidents, a file the engineer can pick up without redoing the diagnosis. The IT department can see what the agent handled and how.
Recurring, documented requests such as password resets and standard access, and known fixes within the assigned scope. Every action is logged.
A fix that affects several users or a shared system, and any configuration change. The agent prepares the intervention, and an engineer approves it before it runs.
A security incident, a critical system or a cause it cannot find. The agent does not intervene: it escalates immediately to the on-call team with everything it has observed.
Several users can no longer connect. The agent groups their tickets, finds the cause in the alerts, applies the documented fix and notifies the people affected.
An employee can no longer log in. The agent verifies their identity following your procedure, unlocks the account and closes the ticket.
An engineer has resolved a new type of incident. The agent turns the resolution into a knowledge base procedure, which is reviewed before publication.
Tickets from your ITSM tool, monitoring alerts, the directory of users and devices, and your knowledge base with its resolution procedures. It also knows the scope the IT department has given it: the systems it may act on and those it must leave to the teams.
It qualifies each ticket, groups requests that belong to the same incident and looks for the cause using alerts and history. If the fix is documented and within its scope, it applies it and checks that the service is back. Otherwise, it routes the ticket to the right team with its diagnosis and what it has already checked.
Resolved, documented tickets, users kept informed of progress, a knowledge base that grows with each new resolution and, for escalated incidents, a file the engineer can pick up without redoing the diagnosis. The IT department can see what the agent handled and how.
Recurring, documented requests such as password resets and standard access, and known fixes within the assigned scope. Every action is logged.
A fix that affects several users or a shared system, and any configuration change. The agent prepares the intervention, and an engineer approves it before it runs.
A security incident, a critical system or a cause it cannot find. The agent does not intervene: it escalates immediately to the on-call team with everything it has observed.
Deployed at Société Générale, the IT Service & Operations Agent handles L1/L2 tickets autonomously — diagnosis, fix, knowledge base — with N2 escalation on complex cases.
30 minutes to validate feasibility on your processes. No slides, no pitch — a concrete conversation.