Security News

ThreatLocker Launches Tool to Keep AI Agents Within Set Boundaries

Australian Cyber Security Magazine · 23 Oct 2026

Key Takeaway Before letting AI agents act inside your business, decide exactly what they may access and run, and enforce those limits with controls outside the AI tool itself.

ThreatLocker has launched Zero Trust AI Control (ZTAIC), a product it says helps organisations use "agentic" AI while limiting what AI agents can access and execute. Agentic AI refers to tools that do more than provide information: they take actions across an organisation's systems. ThreatLocker CEO and co-founder Danny Jenkins said organisations need a way to adopt AI without "handing over the keys to their environments".

The company said ZTAIC applies policy controls outside the AI model or application itself. This responds to concerns that guardrails built into agent workflows may be bypassed or may fail when an agent behaves unexpectedly. Security teams can define what an agent is permitted to do across applications, processes and services. ThreatLocker noted that a single prompt can trigger actions across scripts, browsers, applications and external services, and said ZTAIC is designed to maintain access boundaries throughout those workflows.

ZTAIC includes recommended policies, so security teams can start from pre-built settings and tailor them to their own environment. ThreatLocker said it can be used alongside its existing products, including Allowlisting and Web Content Control, to restrict which AI applications and browser-based services are permitted and what authorised agents can do once running. This summary is based on the opening of the source article only.

AI security agentic AI zero trust ThreatLocker access control

Summarised by CISO AI from Australian Cyber Security Magazine, written with Claude Sonnet 5.5. We link back to every original so you can read it yourself.