In Conversation with Sukhmani Lamba: Building the Future of Work with AI Agents at Microsoft Teams

We sat down with Sukhmani Lamba, Senior Product Manager at Microsoft Teams, to talk about her journey into AI, the future of collaborative agents, patents, platform innovation, and how women in product can thrive in one of the fastest-moving spaces in tech.

Q: What set you on the path to AI and product management?

Sukhmani: I’ve always been drawn to problems where multiple systems—and people—have to work together. Early roles at Deloitte, Wayfair, and Pendo really exposed me to how business growth, user experience, and platform constraints collide in the real world.

When I joined Microsoft, that instinct found the right canvas. Teams is where work actually happens, so building the AI agent platform here means the impact shows up in someone’s life daily — in engineering standups, brainstorms, or sales calls, not just in demos.

And as AI agents enter the workforce as digital coworkers, that impact multiplies. A single agent can now serve thousands of niche workflows across industries.

Q: What’s your approach to product management in AI, especially in enterprise collaboration?

Sukhmani: My approach centers on building human-centric, trustworthy intelligence. Teams is the dominant player in enterprise collaboration, and I focus on defining how humans will interact with AI agents — and how that will reshape organizational structures.

It’s a careful balance between business value, UX, enterprise governance, developer needs, and platform scale.

Q: What have been some of your most impactful areas of ownership at Microsoft Teams?

Sukhmani: I lead parts of our Agent Platform strategy, defining standards that partners and customers use to build AI agents for hundreds of millions of users.

One of the biggest areas I’ve worked on is the roadmap for Microsoft’s flagship collaborative agents like Channel Agent, which acts as a digital project manager — synthesizing conversations and generating status reports.

I also own the Agent Memory charter, defining how agents become contextual and recall information intelligently at enterprise scale — including how memory is shared in collaborative environments.

Another major area was conceptualizing the first API-based Message Extension, which evolved into Microsoft-wide Copilot Plugins and then into low-code Declarative Agents — allowing anyone with an API to spin up an agent quickly.

And I’ve driven the evolution of Adaptive Cards into live, interactive components that support rich media and work across Teams, Outlook, and Loop. Some of this work has been patented.

Q: How do you decide what to build first in a space moving this fast?

Sukhmani: Two main filters:

• Jobs-to-be-done — what are people already hacking together?
• Platform leverage — will one primitive unlock many workflows?

Q: What’s been surprisingly hard about agents in enterprises?

Sukhmani: The first thing is habituation. Agents in the workplace are still new. If you asked me four years ago when we’d start collaborating with digital workers, I would’ve said a decade. But it’s happening now.

The second is trust. Agents in shared spaces must be coherent, contextual, non-noisy, and not hijack conversations, yet not wait forever for perfect instructions. Finding that balance is critical.

Q: You’re active in the product community. Can you tell us more?

Sukhmani: I’ve spoken at Product Coalition, Products That Count, and Products by Women, and I write about how AI is redefining PM roles and digital product design. I also mentor PMs and advise startups in Seattle.

Q: You’ve received a lot of public recognition. What does that mean to you?

Sukhmani: I’ve keynoted at Microsoft Build and Ignite, which together reach 150K+ attendees. I’ve also received patents for several products, joined the Products That Count Advisory Council, and serve as a Product Executive in Residence at Mighty Capital.

Q: What’s your advice for PMs trying to “do AI”?

• Start with a clear decision loop — not “LLM everywhere.”
• Instrument fallbacks and guardrails on day one.
• Ship narrow, then generalize.

Q: Final word for women building in AI and product?

Sukhmani: Pick hard problems. Experiment early. Communicate your impact. And don’t be afraid to pivot — it’s one of the most important skills in a field evolving as fast as AI.