Security | Threat Detection | Cyberattacks | DevSecOps | Compliance

Organisations deploying agents face a challenge: the predominant AI frameworks most organisations rely on do not explicitly address agentic AI. This is true for the big three: ISO 42001, NIST's AI Risk Management Framework (RMF), and the EU AI Act.

AI application security protects AI-powered apps, including those powered by large language models ( LLMs), from unique threats like prompt injection, data poisoning, and model theft. It achieves this by securing the entire lifecycle, including code, data, algorithms, and APIs, using specialized tools and processes that go beyond traditional security measures. It involves securing the AI model’s behavior, training data, and outputs.

See how Torq harnesses AI in your SOC to detect, prioritize, and respond to threats faster. Request a Demo I recently sat between two people who think about the AI SOC operations from completely different angles — and spent 50 minutes watching them land in the same place.

When your CNAPP flags a suspicious dependency in an AI agent container, your WAF logs an unusual API spike, and your SIEM shows a burst of cloud storage calls—are those three separate incidents or one rogue agent attack? Most security teams treat them as three tickets in three queues, investigated by three people who may never connect the dots. By the time someone pieces together that a single compromised agent drove all three signals, the attacker has already moved laterally and exfiltrated data.

Your AI-BOM shows every model, tool, and data source you deployed. But when your SOC investigates an alert about unusual agent behavior, that inventory tells them nothing about what actually happened at runtime. Static AI-BOMs document what you intended to run. Attackers exploit what your AI workloads actually do in production: which APIs they call, what data they touch, and how they use approved tools in unapproved ways.

Fresh off two weeks of back-to-back meetings in Washington, DC, and on the floor/in the wings of the RSA Conference, one theme echoed through nearly every conversation I had with senior government officials and public policy leaders from global technology companies: agentic AI security is the defining emerging security challenge of this moment — and policy is not keeping pace.

In 2025, we embarked on a new journey to secure the most important technology transformation of this decade – generative AI. Our vision is to help companies secure their AI fast, so that they can innovate on the cutting edge and put AI and agentic use cases into production. To do this, we built Evo, the world’s first agentic orchestrator for AI security. The foundation of any product is customer needs.

“AI-powered DAST” is everywhere. It signals progress, but assumes something fundamental was missing. It wasn’t. DAST struggled not from lack of intelligence, but from lack of depth. Most tools never reached inside authenticated, stateful, multi-step journeys where real logic, sensitive data, and critical vulnerabilities exist. That’s the part Appknox solved years ago. AI here is not a reset. It is an accelerator, applied to a system already operating where risk actually lives.

Agentic AI workloads are shipping to production on Kubernetes faster than the standards to secure them. Many teams deploying autonomous, tool-calling agents as containerized microservices do so without a shared baseline for securing or monitoring those containers. The CNCF AI Technical Community Group recently published a comprehensive article on cloud-native agentic standards, marking the first attempt to define best practices for such deployments.