Most AI agent conversations start with what the agent can do. Very few focus on how you manage, monitor, and trust those agents once they're in production. At Cisco Live, I sat down with Kamal Hathi, SVP & GM of Splunk at Cisco on The Ravit Show, to discuss what enterprises need beyond AI models to make agents reliable, secure, and trustworthy.

A few key takeaways from our conversation:

* Moving AI agents from demos to production requires visibility into how they operate, make decisions, and interact with enterprise systems.

* As organizations deploy more agents, observability becomes critical. Without it, AI can quickly become a black box.

* Data remains one of the biggest challenges. Enterprises are looking for ways to reduce tool sprawl while maintaining a unified view across their environments.

* Security and observability are no longer separate conversations. The faster teams can connect operational issues with security events, the faster they can respond.

* Making AI accessible is important, but governance cannot be an afterthought. Innovation and control must go hand in hand.

One thing that stood out to me:

The future of AI isn't just about building smarter agents. It's about creating the trust, visibility, and governance needed to operate them at enterprise scale. Great conversation with Kamal on the next phase of enterprise AI and the role observability will play in making it successful.

#data #cisco #ciscolive #ai #theravitshow