AI Agent & Copilot Podcast: Donna Sarkar of Microsoft on Moving AI Agents from Experimentation to Production

In this episode of the AI Agent & Copilot Podcast, John Siefert, CEO of Dynamic Communities and Cloud Wars, sits down with Dona Sarkar, Chief Troublemaker, Enterprise AI Advocacy at Microsoft, to explore what it really takes to move AI agents and copilots from experimentation into production. Their conversation previews Sarkar’s keynote at the 2026 AI Agent & Copilot Summit NA and dives into practical adoption, human-centered AI, and lessons learned from real-world enterprise deployments.

Key Takeaways

• Enterprise advocacy bridges the gap: Sarkar explains that enterprise cloud advocacy exists to translate Microsoft product capabilities into practical, real-world business solutions. Rather than selling tools, her team focuses on enablement — creating demos, workshops, and labs that show how AI agents, Copilot Studio, Azure, and Power Platform can actually be deployed inside organizations.
• Production is harder than experimentation: Building an AI agent is easy; deploying it responsibly is not. Enterprises struggle with permissions, ownership, data readiness, and governance once agents move into production. These challenges reveal why successful AI adoption requires cross-functional collaboration between IT, business units, and governance teams.
• Not all work should be automated: Sarkar cautions against replacing meaningful human interactions with automation simply because it’s possible. Instead, organizations should focus AI on prioritization, analysis, and repetitive tasks — freeing humans to spend more time on creativity, judgment, and relationship-building. “We really need to go draw a big old line in the sand and say, these should be uniquely human to human activities," she says. "These should be uniquely AI to human activities. These should be uniquely AI to AI activities.”
• Human connection matters more than ever: Despite fears that AI would reduce in-person interaction, both speakers observe the opposite trend. Conferences and professional gatherings are thriving because people crave perspective, not just information. While AI can surface data instantly, point of view comes from lived experience.
• Failure is part of responsible AI adoption: Sarkar openly shares that "The number of agents I’ve had to take down is probably like 50% of the agents I built.” These failures weren’t wasted effort; they informed better tooling, clearer governance, and improved workflows. Microsoft’s rapid release of new AI tools reflects lessons learned internally before being shared with customers.

Visit Cloud Wars for more.