AI Pulse · ANZ AI Watch

A Jamf-commissioned survey of 687 ANZ IT and security leaders points to a problem many businesses are only starting to see. AI use is spreading quickly, with 72.9% of respondents saying they have deployed it in some form. Among Apple-first organisations, 81.7% reported or expected an AI-related security or cost incident.

The more useful figure is the gap between experimentation and deep adoption. Organisations with deep AI integration were 40% more likely to report an incident than those still exploring AI, 27.1% versus 19.4%. That does not mean AI itself is the problem. It suggests the risk rises when staff begin using local AI tools, connecting them to business information, or adopting new workflows without clear visibility and controls.

This is vendor-sourced survey data, not independently verified, so treat it as a signal rather than a final verdict. Still, the underlying issue is practical. Traditional security tools may not show which AI applications employees are using, what data is entering those tools, or where unexpected costs are building up.

For an owner or operator, the immediate job is not to block every AI tool. It’s to establish a clear approved-tool list, define what data can and cannot be entered into AI systems, and give one team responsibility for monitoring usage, access, and spend. This is the kind of thing we build into an AI command centre, because AI governance needs to be part of normal operating control, not a once-a-year policy document.