This week’s Modern Work Mondays episode comes to you from San Francisco, where Nolan and Jeff are on the ground at Microsoft Ignite. With more than twenty thousand attendees, the event is filled with conversations about what’s next in modern work. One theme stands out across those conversations: AI agents are growing quickly, and organizations are now figuring out how to manage them in a consistent, scalable way.

Getting control of the agent ecosystem

As more teams build agents through low code tools or internal development, the need for structure becomes clear. Organizations are trying to understand how to manage these agents throughout their lifecycle, including identity, access, ownership, and what happens when the creator of an agent leaves the company. Interest in agents is rising quickly, but without clear governance, it becomes difficult to grow their use with confidence.

Experimentation, sandboxes, and skills

Alongside this shift, teams are experimenting more. Low code and no code tools make it easy to explore ideas, whether that means dropping files into SharePoint or testing concepts in Copilot or Foundry. These early trials help people see what is possible, but turning an experiment into something production ready requires real resources and often the support of a systems integrator.

Sandboxes play an important role in that early phase. They give teams space to see what is real and what is worth investing in. At the same time, organizations are putting more focus on skilling so people understand how these tools work and what they can access. With that foundation in place, teams can start applying AI to their own areas of expertise in ways that deliver meaningful value.

Understanding latency in voice interactions

Another topic gaining attention at Ignite is the role of latency in voice experiences. Modern apps give users a visible cue while something is processing, but voice does not. Silence on a radio feels longer than it is, and even a short delay can make people question whether anything is happening. In highly distributed environments, the most frequent voice interactions need to run locally to feel immediate, while heavier workloads can stay in the cloud. When latency stays low, voice fits naturally into daily work. When it does not, frontline teams lose patience quickly.

Turning potential into practice

Across all of these areas, a common thread emerges. AI is becoming part of everyday operations, and its value depends on stability, clarity, and real-world performance. Experimentation helps people get started, but long-term impact requires strong governance, solid skilling, and technology that works reliably in the moment. As organizations continue exploring agents, low code tools, and voice, the focus remains on creating solutions that fit naturally into how people work.

Watch this week's episode: