Change Request Review Agent · Use case
How can a telecom operator use an AI agent to review a network change request on its own infrastructure?
A change request review agent checks a request to change the network against the operator's own change rules. It reads the method of procedure, the rollback plan and the window, looks for other changes planned on the same equipment, and lists what is missing or in conflict. The change board approves every change and an engineer makes it. The agent approves none and makes none.
The request looked complete.
The conflict was found on the night.
- ✗ 01A reviewer checks dozens of requests a week against rules held in memory and in a long document.
- ✗ 02Two changes on the same equipment in the same window are raised by different teams, and nothing puts them side by side.
Where it stops for a person
Always. The change board approves, and an engineer makes the change.Always. The change board approves every change, and an engineer makes it through your change process.
It reviews the request.
People change the network.
Step · 01
A change manager hands the agent a change request.
Step · 02
It checks the request against your change rules: a method of procedure, a rollback plan, a window, a named owner.
Step · 03
It reads the inventory for the services and customers that depend on the equipment.
Step · 04
It reads the change calendar for other work on the same equipment or the same route.
Step · 05
It writes the review: what is missing and what is in conflict, each with the rule or the record behind it. The change board decides.
Reads from · connectors
Your own APIs
The change request and the calendar, through your change system's own API. The inventory, through its own API; where it follows a TM Forum Open API, the description is already public. Your change rules are a knowledge source.
Acts through · MCP servers
Nothing. It has no tool that reaches network equipment, and none that approves or schedules a change.
What you set up
- Connect the change system and the inventory with read-only access.
- Index your change rules.
- Test against past changes.
Where it runs
The agent, its knowledge store and the model run on infrastructure you own or control. You choose the model and serve it yourself, so questions, documents and answers are not sent to a public AI service. Change records and network inventory stay inside the operator.
How you get it today
This page is labelled Use case: the agent is built in Katonic Studio. You describe the job, connect what is named above, add test questions and release the agent once it passes them. When the Catalog holds a tested template for this job, this page will carry the Template label and show how to take it.
Can the agent approve a low-risk change by itself?
No. It has no tool that approves. Which changes need the board is your rule.
Does it check the commands in the method of procedure?
It checks that the procedure covers what your rules require and lists what it could not find. It does not run or test a command.
What if the inventory is wrong?
The review names the records it read. A service missing from the inventory is missing from the review.