Agent Hub

A unified command center that helps enterprise admins adopt AI safely, confidently, and at scale.

Oct 1, 2025

Role

Product Designer

Timeline

Jan 2025 – Ongoing

As Microsoft moved toward a fully autonomous contact center vision, customers needed a better way to understand and scale AI features like Copilot and Autonomous Agents. Agent Hub was created as a centralized experience to guide enterprise admins through the journey of learning, adopting, and measuring AI with confidence.

Problem

Enterprise admins were eager to deploy new AI tools, but faced major roadblocks. Documentation was scattered, governance reviews were complex, and the rollout process was unclear.

Without a single place to evaluate, test, and measure success, adoption stalled. Admins were forced to rely on manual processes, disconnected resources, and repeated privacy reviews.

We needed to build a centralized hub that could support a wide range of users, from cautious first-time adopters to experienced IT teams, and reduce friction at every stage of the journey.

Research and Insight

Research and Insights

To get started, I gathered legacy work and reframed the problem with input from multiple PMs, including a new lead I helped onboard. We conducted competitive research and found that many platforms already offered their own versions of a hub, centralizing transparency, controls, and guidance.

Meanwhile, we had answers to many customer concerns already, they were just spread across disconnected sources.

We identified four core needs: Learn, Adopt, Track, and Verify. To move quickly, we scoped Verify out of this phase and focused on the first three pillars, with Track later renamed to Measure. This gave the team a shared structure and strategic direction.

To get started, I gathered legacy work and reframed the problem with input from multiple PMs, including a new lead I helped onboard. We conducted competitive research and found that many platforms already offered their own versions of a hub, centralizing transparency, controls, and guidance.

Meanwhile, we had answers to many customer concerns already, they were just spread across disconnected sources.

We identified four core needs: Learn, Adopt, Track, and Verify. To move quickly, we scoped Verify out of this phase and focused on the first three pillars, with Track later renamed to Measure. This gave the team a shared structure and strategic direction.

Design

Agent Hub was structured into three key experiences: Learn, Adopt, and Measure.

I led the design of both Learn and Adopt.

Learn
We created a centralized Learning Hub within the Copilot Service Admin Center. I focused on reducing friction by making technical concepts like data flow, compliance, and governance easier to understand and explain. This helped unblock leadership approval and privacy reviews.

Adopt
This was the most complex part of the project. The Rollout Manager needed to guide customers through a phased activation of AI features while integrating with existing systems.
I designed for clarity, trust, and step-by-step progression, making the flow intuitive even for first-time users.

We faced real engineering constraints that shaped what was possible. I worked closely with developers to adjust designs without compromising the user experience.

Measure
Although I was not the primary designer on this pillar, I helped ensure the experience aligned with the broader structure of Agent Hub. It allows admins to track usage, gather feedback, and report impact over time.

Agent Hub was structured into three key experiences: Learn, Adopt, and Measure.

I led the design of both Learn and Adopt.

Learn
We created a centralized Learning Hub within the Copilot Service Admin Center. I focused on reducing friction by making technical concepts like data flow, compliance, and governance easier to understand and explain. This helped unblock leadership approval and privacy reviews.

Adopt
This was the most complex part of the project. The Rollout Manager needed to guide customers through a phased activation of AI features while integrating with existing systems.
I designed for clarity, trust, and step-by-step progression, making the flow intuitive even for first-time users.

We faced real engineering constraints that shaped what was possible. I worked closely with developers to adjust designs without compromising the user experience.

Measure
Although I was not the primary designer on this pillar, I helped ensure the experience aligned with the broader structure of Agent Hub. It allows admins to track usage, gather feedback, and report impact over time.

Agent Hub was structured into three key experiences: Learn, Adopt, and Measure.

I led the design of both Learn and Adopt.

Learn
We created a centralized Learning Hub within the Copilot Service Admin Center. I focused on reducing friction by making technical concepts like data flow, compliance, and governance easier to understand and explain. This helped unblock leadership approval and privacy reviews.

Adopt
This was the most complex part of the project. The Rollout Manager needed to guide customers through a phased activation of AI features while integrating with existing systems.
I designed for clarity, trust, and step-by-step progression, making the flow intuitive even for first-time users.

We faced real engineering constraints that shaped what was possible. I worked closely with developers to adjust designs without compromising the user experience.

Measure
Although I was not the primary designer on this pillar, I helped ensure the experience aligned with the broader structure of Agent Hub. It allows admins to track usage, gather feedback, and report impact over time.

Impact

Projects

Agent Hub quickly became a strategic priority. After demoing the vision to leadership, the project was greenlit for release.

Outcome highlights

• Learn hub on track for General Availability in October 2025
• Rollout Manager entering Public Preview in October 2025
• Demo at leadership offsite led to immediate prioritization
• Adopt experience was called out by leadership as a standout moment
• Helped reduce delays in AI deployment by centralizing governance resources

As Microsoft moved toward a fully autonomous contact center vision, customers needed a better way to understand and scale AI features like Copilot and Autonomous Agents. Agent Hub was created as a centralized experience to guide enterprise admins through the journey of learning, adopting, and measuring AI with confidence.