David Pidsley from Gartner posted on LinkedIn: “Centralizing AI agent management under IT could stifle adaptive governance and innovation by focusing too much on operational “how” rather than strategic “why.” Instead, decision-making authority should remain distributed across business units, ensuring alignment with customer needs and strategic goals.“

“AI agents should not be viewed as “workers” or “interns” but rather as high-agency decision systems—designed to automate or augment specific decision model, executed and monitored for self-improvement.”

“A design principle for AI agents should be decision-centricity. AI agents are more akin to “the new apps” than human employees, and “managing” them requires a product management approach, not a talent management framework.” Link