For MSPs evaluating AI on client tickets
AI agents for MSPs, working on the device
“NetZen seems much more aligned with where support operations are heading rather than just adding an AI integration onto an older model.”
Thanksgiving Offer
$50$100USD per technician per month*
Prefer a walkthrough? Book a demo
What does an AI agent do on a ticket that a chatbot does not?
A client's app will not open (an example, not a customer). A chatbot writes back a suggestion and your technician still has to inspect the device. In NetZen's service desk, the AI collects the details your intake policy asks for, then checks the connected machine through the automation packs you enabled: diagnostics, networking, printers, software, Microsoft apps and patching. Microsoft 365 and Google Workspace checks are read-only.
When a tested automation matches, the run waits for the approver you set for it: Requester, IT Support, Site Manager or Not Required. Every run lands in History with its status, who started it and the output per device.
When the request needs a person, the AI stops. The ticket moves to Waiting for agent with the findings and logs attached, and Live progress shows what happened, the next step and its owner. The RMM and automations it works through are the ones your technicians use; one run book serves every client it is scoped to.
Who decides what the agent is allowed to do?
You do, per automation and per client. Each pack is switched on for the clients you choose, each automation has a permission (end user, support agent or both) and an approver, and custom automations start disabled until your technician has tested them. NetZen is the full PSA, priced per technician.
Before you install anything
A chatbot writes a reply. An AI agent acts on the ticket: it asks for the details your intake policy needs, runs checks on the connected device, proposes a matching fix and runs it once an approver says yes.
Changes started from tickets follow the approvers you set per automation, and the AI cannot approve its own changes. A ticket from an end user can read only that user’s own account.
No. It does the intake and the device checks, then either runs an approved fix or hands the ticket to a technician with its findings and logs. Tickets that need a person land in Waiting for agent with the next step and its owner visible.
US$100 per technician per month. Endpoints, clients and end users are not billed, so every client device the technician looks after is included. The 30-day trial needs a work account, not a card. NetZen is hosted on Google Cloud in the US.
What did an MSP that wanted more than a chatbot report?
Cinema Hero Solutions, an MSP handling about 500 tickets a month by its own count, was not looking for a chatbot: it wanted questions that uncover the problem and a starting point for technicians. Matt Kerekes, its CEO, on the handoff:
“It hands the ticket to us and says, here are the details, here are the logs, now you can dive into it, rather than spending 30 minutes to an hour trying to find out what is actually happening.”
The customer's own words, not a measured result. Read the customer interview.
How do you test an AI agent on your own tickets?
Start the 30-day trial with a work account, no card. Connect your computer, try the sample app request and follow its approval and device action, then put a real client workflow through the evaluation guide.