Security | Threat Detection | Cyberattacks | DevSecOps | Compliance

Let’s face it, AI is everywhere now. It has moved from novelty to necessity, reshaping the way we work, make decisions and secure our organizations. It guides how we plan trips, shop for essentials and discover information – but one of its most profound impacts is happening across enterprise environments.

In February 2023, a Stanford University student conducted a study that turned into one of the most widely followed security tests in AI history. Kevin Liu performed a simple prompt-injection attack, tricking Microsoft Bing Chat into disclosing its internal codename, Sydney, and exposing the entire list of its system prompts. The attack utilized no high-end toolkit, no zero-day, and no privileges, only specially crafted natural language.

We recently published a series of polls across our social channels to get a pulse on some of today’s application security concerns with AI. These recent conversations with our community reveal a clear and urgent shift in the application security landscape. Results show that while established challenges like software supply chain security remain top of mind, the rapid pace of AI has created a new center of gravity for anxiety.

SOAR was created to help security teams work faster and more consistently by automating and orchestrating core security operations. It has always had to adapt to new and evolving technologies, but our current AI era has brought about a turning point. As cloud environments scale, manual playbooks can’t keep up. Now, it’s not enough to automate. We need systems that can understand the context they’re running in and adapt accordingly.

Agentic browsers are quickly becoming part of everyday work. Tools like ATLAS, Comet, and Dia can read web content, navigate SaaS tools, interpret instructions, and act on behalf of a user. They promise faster execution and higher productivity but they also introduce new risks that traditional security tools are not designed to see. As these browser-based agents spread across both managed and unmanaged devices, the enterprise attack surface grows in ways that most teams can’t quantify.

AI agents are increasingly connecting to more systems and workflows. They read structured data, follow multi-step instructions, and can reach deep into applications and developer environments. The same capabilities that make them powerful also create new opportunities for attackers. As Zenity Labs continued to study these emerging attack classes, we noticed a pattern starting to appear.

Data is never the problem. Security teams rarely complain about having too much of it. The real danger comes from data that sits unconnected and unexplained. What teams actually need is data that is actionable and converges into meaning. Data that cuts deeper than surface level signals. Data that reveals what is unfolding and what needs to happen next.

The rapid adoption of AI has introduced a new, semantic attack vector that many organizations are ill-prepared to defend against: prompt injection. While many security teams understand the threat of direct prompt injection attacks against AI agents developed by their organizations, another more subtle threat lurks in the shadows: indirect prompt injection attacks.