ControlUp has announced Pulse AI Resolve, an AI tool that investigates and can fix IT problems on employee devices, with administrators deciding how much it may do on its own. The useful wrinkle is that the system is designed to move routine fixes closer to the problem while escalating harder cases with diagnostic context already gathered.
Watch Desk analysis
What happened
Pulse AI Resolve monitors endpoint performance, can investigate live device data and determine a likely cause. An employee can describe a problem in plain language, or an IT administrator can send in a ticket with little context. The system can also detect issues proactively, according to ControlUp’s announcement, carried by Help Net Security.
ControlUp describes four levels: matching known patterns to approved fixes; assembling actions from a library of more than 400 custom options; using agentic AI to reason through more complex problems; and handing cases to a person with diagnostic information attached. The first three are designed to work within policies set by IT. Administrators can require approval for system changes, restrict which device groups are investigated, and set rules for escalation.
Why it matters
The pitch is not simply that AI can answer a support request. It is that the software can investigate a device, take some actions and, in certain cases, coordinate changes across connected systems. That could resolve routine issues before they become help-desk tickets, while giving staff a clearer handover when automation reaches its limits.
The practical question is where an organisation draws those limits. ControlUp says diagnostics are performed locally using targeted, time-bounded data, and that the tool stops and escalates when it reaches the boundaries set by IT. Those are details of the company’s announcement, not independent evidence of how the system performs in deployments.
Our read
A tiered support model is a more persuasive idea than giving an agent a broad set of permissions and hoping the guardrails hold. Here, administrators are meant to choose which tasks can run automatically, which need approval and when a person takes over. That puts a real operational decision in front of IT teams, rather than pretending every ticket deserves the same level of autonomy.
The test will be whether the controls are understandable and useful under pressure. Four tiers sound tidy on paper; a messy device problem has a habit of ignoring the diagram.
What to watch
- Whether organisations can set clear, practical rules for automatic fixes and approvals.
- How accurately the system diagnoses problems using endpoint data.
- Whether the 400-plus-action library and multi-system integrations work reliably in real support environments.
- How often complex cases reach a human, and whether the gathered context makes that handover better.
Discussion spark: Which IT fixes, if any, should an AI system be allowed to apply automatically, and which should always require a person’s approval?
Sources and evidence
- Source update (28 October 2026, 14:46 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.