Control Plane for Agent Memory - Permissions

Join Vas Markovic, Founder and CEO at cognee, for a practical look at the question every team runs into once agents start sharing memory: who is allowed to see, change and share what? As soon as more than one agent, team or customer works on the same memory, permissions stop being an afterthought. They are a necessary pre-condition for scaling your agentic system across the entire organization. In this session, Vas will walk you through how cognee handles permissions for agent memory, make company data accessible to your entire team, all while while keeping sensitive data secure and where it belongs. We'll cover how cognee is designed to: Separate memory by team and tenant: Keep each team's, customer's or project's knowledge in its own space, so agents only reference what they are meant to see. Control read, write and share access: Decide per user and per agent who can read a dataset, who can add to it, and who can share it with others. Keep humans and agents in the loop safely: Let agents act on company knowledge without exposing data across teams, customers or roles. Vas will close with the practical side: how to design your first permission model, which boundaries to set on day one, and the mistakes that most commonly lead to leaking or siloed memory. 📅 November 4, 9:00 - 10:00 AM PT (12:00 - 1:00 PM ET / 18:00 - 19:00 CET) · Register For: developers, technical leads, and teams running AI agents on shared company data.