Niten Aggarwal

Talk Title:

When AI Agents Start Taking Action: Rethinking Security, Trust, and Governance for Enterprise AI

Abstract: Enterprise AI is moving beyond answering questions and generating content. Agents can now reason through a request, discover available tools, decide what to invoke, and take actions across enterprise systems. This creates a different challenge from the API-driven applications we have been building and securing for years. Once an agent starts acting on behalf of a user, who is it really acting for, what should it be allowed to do, and how do we control and trace those actions?

In this talk, I will discuss the architecture considerations behind this shift, including how identity and authorization should carry from the user to the agent and the tools it invokes, how MCP-based tools can be discovered and governed, and where traditional API controls end and agent-specific governance begins. I will also discuss reliability when agents perform state-changing actions and the importance of maintaining an audit trail from user intent through agent decision, tool invocation, and the resulting enterprise action.

Biography: Niten Aggarwal is an enterprise architect and expert software engineer with more than 20 years of experience designing and modernizing large-scale enterprise systems. His work spans distributed cloud architecture, multi-region resiliency, API and workload security, enterprise data platforms, and AI-enabled systems. Over the course of his career, he has worked across telecommunications, healthcare, retail, life sciences, and digital commerce.