How Leading Teams Discover, Observe, and Secure Every Agent in Production
ABOUT EVENT
The co-founders of Natoma (now Snowflake) and Trust3 AI invite you to an educational session in which you will learn how to responsibly accelerate the deployment of AI systems for your organization.
Most enterprises are running 3–10× more AI agents than their teams realize. For practitioners, that means a rapidly expanding attack surface that's difficult to discover, govern, and secure.
Every agent touches data, takes actions, and may operate as part of Shadow AI. The critical questions are: Which agents are running? What can they access? What are they actually doing? And who is accountable when one goes off script?
This session explores how you can discover every agent running in production, secure every action those agents take on enterprise data, and responsibly accelerate AI adoption.
Expect candid talks, a live end-to-end demo, and direct conversation with the people building the security and observability layer for the agentic enterprise.
SPEAKERS
Paresh Bhaya, Co-founder, Natoma (now Snowflake). Co-built the enterprise MCP gateway Snowflake acquired to govern how agents connect, authenticate, and act across systems.
Neeraj Sabharwal, Co-founder, Trust3 AI. Helps enterprises safely scale AI adoption with strong data governance and access control.
Matthew Sullivant, Head of Product, Trust3 AI. Leads AgentDOS, the control plane for discovering, observing, and securing agents across any framework and cloud.
Mark Castillo, Founding AI Solutions Architect, Trust3 AI. Partners with enterprises to design and deploy agentic governance systems in production.
Trust3 AI, host and live demo team.
WHAT TO EXPECT
5:00 to 6:00 PM | Welcome Reception & Networking
Check in, grab a drink, and enjoy light bites as guests arrive. Connect with fellow data and AI practitioners & pioneers before the program begins.
6:00 to 6:05 PM | Welcome & the Agentic Reality Check (Trust3 AI)
Enterprise AI is scaling; governance isn't. A quick look at shadow AI, rogue agents, and why this lands squarely on both the security team and the AI budget.
6:05 to 6:20 PM | Securing the Agentic Enterprise (Paresh Bhaya, Natoma, now Snowflake)
Governing agents that act on data across Snowflake Intelligence and Coco: the Natoma MCP gateway, identity, and purpose-bound access, built from the data layer up.
6:20 to 6:40 PM | AgentDOS: Discover, Observe, Secure (Matthew Sullivant, Trust3 AI)
Finding the agents you're undercounting by 3 to 10x, tracing every prompt and tool call, token-level observability, and turning agent risk and spend into measurable AI ROI. Security doesn't stop at the MCP layer, it has to reach the data itself. A first look at the Policy Agent, Trust3's approach to fine-grained access control at the data layer.
6:40 to 7:00 PM | Live Demo: Policy Agent in Action (Trust3 AI)
A live walkthrough on Snowflake: author an access policy in natural language, watch Trust3's Policy Agent generate and apply the underlying roles and grants automatically, and see that same policy govern what an agent is allowed to touch.
7:00 to 8:00 PM | Q&A + Open Networking
Open Q&A with all speakers, followed by continued conversation over drinks.
WHO SHOULD ATTEND
Data and AI practitioners & pioneers
LOCATION
Snowflake Office, 135 Constitution Dr, Menlo Park, CA 94025