Agentic AI is only interesting when it leaves the lab and takes responsibility for real outcomes. In this episode, I reconnect with Raj Koneru, Founder and CEO of Kore.ai, to talk about what that shift looks like inside large enterprises. Raj has been building conversational systems long before chatbots became dinner table conversation, and he is clear about where the action is now. Understanding intent is table stakes. The next frontier is planning, reasoning, and executing tasks through agents that can work across departments, respect policies, and prove their value in minutes, not months.

We walk through how Kore.ai frames the stack. There is a layer below the line, where chips, data centers, clouds, and model providers keep advancing at speed. Above the line is where business value shows up. That is where companies design, deploy, and manage agents for customer service, employee productivity, and process automation. Raj describes Kore.ai as the operating system for that upper layer. The goal is simple to say and hard to deliver. Let teams build agents without writing code, plug into whichever models and clouds they prefer, and keep control through governance, security, and measurement.

Interoperability runs through this conversation. Kore.ai partners with Microsoft, AWS, and G42’s Inception so customers can pick the environment that suits them. Vendor lock-in is a real anxiety for leaders. Raj’s answer is to be model agnostic, data agnostic, and cloud agnostic, while conforming to open standards. That way, enterprises spend their energy on the agents themselves rather than on maintaining a platform. It is a pragmatic view that reflects what I keep hearing on conference floors.

Scale brings its own lessons. At Kore.ai’s volume, latency, accuracy, and security are non-negotiable, and governance is a daily practice rather than a slide in a deck. Raj talks candidly about no-code democratization, where a business user can assemble an agent for a focused task, then graduate to more complex workflows when ready. We also touch on the rise of on-device models in smartphones and where the boundary sits between quick local tasks and heavier actions that still rely on hosted models.

If you care about agentic AI that does real work, this one is worth your time. Raj shares a thoughtful take on collaboration, standards, and why leaders should read widely, question everything, and start building with what is available today. Links are in the show notes.