Rameshkumar Varanganti
From API-Centric Integration to Agentic AI: Building Secure, Intelligent Enterprise Systems with Model Context Protocol (MCP)
Abstract:
Enterprise APIs typically support applications that follow predefined workflows. AI agents can select tools and take actions based on the information they receive, raising important questions about permissions, approvals, and accountability. This keynote explores how Model Context Protocol (MCP) can expose enterprise capabilities as tools for AI agents, alongside the controls needed to govern their use. Through an integration incident example, the session will show how an agent can gather evidence, propose a recovery action, and request approval before retrying a failed transaction. It will also examine authentication, scoped permissions, human oversight, and audit logging, and discuss how teams can assess whether agent-assisted workflows improve incident response while maintaining operational control.
Profile:
Rameshkumar Varanganti is an enterprise integration lead, independent researcher, and IEEE Senior Member with over 20 years of experience designing and delivering enterprise integrations. His work spans SAP Integration Suite, API management, and cloud integration across HR, finance, and supply chain systems. His research focuses on agentic AI, secure tool access through the Model Context Protocol (MCP), and governance of AI-enabled enterprise workflows. He created Agentic Integration Hub, a public resource where he shares practical guidance on enterprise AI agents, MCP, and API integration.