Those things we swore we'd never do again after the cloud computing migration? Fact is, we are repeating some of those same missteps now in the agentic AI era.

In this episode of the Enterprise Edge® Podcast, Mark Vigoroso sits down with Sujata Joshi, Director of SolutionDelivery at Cisco, and Sridhar Balasubramanian, VP of Data and AI at CloudPaths, to name mistakes like lift-and-shift, reborn as AI agents bolted onto broken processes.

The stakes are real, with a Barclays survey showing 86% of CIOs plan to pull workloads back off the public cloud and Gartner predicting 40% of enterprises will cancel or demote autonomous agents by 2027. The two enterprise tech veterans in our studio lay out what to do instead:

• Pilot on your ugliest data, not your happy path.
• Stop driving a Ferrari through rush-hour traffic when plain automation will do.
• Keep agents out of requirements discovery until the tribal knowledge is written down.
• Replace the change advisory board (CAB), already dead at 50 changes an hour, with automated guardrails.

Our guests’ verdict: the bottleneck is no longer capability but governance, and the payoff is revenue, not just productivity. The one cloud-era move worth copying outright is a shared responsibility model put in writing before the incident, "so that when something burns, there is an artifact to point to instead of an argument."

Stream this practical episode now - before your next AI agent goes live - and be sure to LIKE, SHARE, and COMMENT!