When teams were building RAG pipelines and making direct API calls, a proxy was enough. That era is over. Agents that plan, call tools, and act autonomously introduce risk that traditional proxies, AI firewalls, and DLP solutions were never designed to handle.

‍

The problem is no longer about inspecting data moving in and out. It is about understanding behavior.

‍

In this webinar, James Berthoty from Latio Tech, Ophir Dror CPO at Lasso, and Or Oxenberg AI Security Engineering Team Lead at Lasso, introduce an intent-based framework for agent security.

‍

The panel will walk through two approaches to putting it into practice: SDK and hooks-based integration that embeds security directly into the agent lifecycle, and eBPF-based runtime sensing that operates at the infrastructure level. We will cover what each gives you, where each falls short, and how they compare to traditional approaches. The session closes with a live demo of Lasso in action.

‍

You'll walk away with:
-A practical threat model for agentic systems at build time and runtime
-A clear comparison of intent-based security versus traditional proxy and -AI firewall approaches
-A live look at Lasso securing real agentic workflows

‍