Zscaler pitches AI agents and an 'agentic SOC' as the next growth engine for its Zero Trust business

By Zendoric · September 13, 2026. Zscaler, the cybersecurity company that trades on the Nasdaq as ZS and that popularized the Zero Trust cloud security model, put AI agents and what it calls an 'agentic SOC' at the center of its latest conference, according to Yahoo Finance. The headline is explicit in its message —these technologies 'fuel the growth ambitions' of the company around Zero Trust— but the available note offers no adoption figures, named customers, prices or a concrete product roadmap. With this material, the honest thing to do is stick to what can be stated without speculating.

A SOC (security operations center) is the team and the tools that monitor a company's network to detect and respond to incidents. Talk of an 'agentic SOC' refers to AI agents —programs capable of investigating alerts, correlating scattered signals and proposing or executing responses more autonomously than a simple chat assistant— taking on part of that work, today swamped by alerts that human analysts cannot keep up with. Zero Trust, Zscaler's core business, is the principle of not trusting any user or device by default, inside or outside the corporate network, and continuously verifying every access; Zscaler sells it as a cloud service through its Zero Trust Exchange platform.

In general, cybersecurity is one of the sectors where the discourse on agentic AI has moved faster than public evidence of large-scale deployments, and this is not the first company in the sector to use a customer conference to reinforce its growth narrative for investors. That does not invalidate the underlying idea: security alert triage is exactly the kind of repetitive, noise-saturated task that, as we have been pointing out in our analysis of AI and employment by sector, is the most exposed to automation when expert judgment remains scarce.

Our reading is that this announcement should be treated for what it is: a piece of corporate narrative at a company event, not a milestone of demonstrated capability. The question that really matters —and that this material does not allow us to answer— is how much real autonomy those agents exercise in production versus how much is still assistance supervised by human analysts. If the 'agentic SOC' delivers on its promise, the underlying benefit fits our long-term thesis: cybersecurity carries a chronic shortage of specialists, and freeing those analysts from routine alerts so they can concentrate on higher-judgment decisions would be a small but concrete example of abundance applied to a sector that is currently short on talent.