The moment your LLM starts creating code or making decisions, it’s no longer just a tool — it’s a user. And most orgs haven’t built security models to handle that. Agentic AI breaks traditional assumptions across the board:

• AppSec must validate code not just written by devs, but by models
• Data Security must control for unintended access through AI-driven workflows
• SecOps must detect non-human activity without drowning in false positives

This keynote walks through how OpenText is applying identity-first principles to protect against a new breed of risk: one that’s fast, scalable, and doesn’t ask permission.