My role
Lead UX Designer
Timeline
8 weeks
Team
1 designer · 1 PM · 6 engineers
Focus
Data visualization
Work is enterprise focused, therefore not many images of work may be published publicly. Low-fidelity images provided when applicable.
Quick summary
I contributed to the analytics experience supporting Dragon Copilot in Dragon Admin Center, helping healthcare administrators understand adoption, engagement, license utilization, and where action/intervention may be needed.
overview
Dragon Copilot introduced powerful AI-assisted workflows for healthcare organizations, but administrators needed a way to understand how those capabilities were being adopted across their organization. The analytics experience in Dragon Admin Center created a customer-facing view into adoption, engagement, license health, and usage patterns for Dragon Copilot.
My contribution focused on the analytics layer that supports Dragon Copilot. I needed to find what kind of data do administrators need in Dragon Admin Center to make informed decisions on licensing, usage, and adoption, and how can we represent that data to them in a scannable way? In addition, what can we automate to ease the user’s burden?
The challenge
With the Dragon Admin Center platform finally having a framework for analytics, it was time to put it to the test. Organizations should be able to understand if they are getting their money’s worth with the AI licenses they are purchasing. We can provide raw API data for them, but for non-technical roles, they need to be able to look at their data and make quick decisions based on if users are returning and if engagement is weak. If a specific location within the healthcare organization isn’t using its licenses and tooling, the organization will want to know why and how they can either save money on licenses or help physicians utilize Dragon Copilot.
Prior to this, there was no way to understand license usage and feature usage. Through user interviews, it was discovered that administrators not only want to see that information but be able to understand what kind of actions can be taken and where to improve the overall experience on Dragon Copilot.
The design question
How do we visualize complex data for easy understanding?
my role
My role was to help shape the analytics experience and provide design guidance for how Dragon Copilot reporting should live inside Dragon Admin Center. What should the report’s layout, style, and intent be? What information is the most useful to admins? What information will result in action from the admins? These guidelines and decisions will also need to be reflective for the Dragon Copilot suite of products. After providing guidance, the Dragon Copilot team should be able to iterate on their work without needing me to be directly involved.
design evolution
The goal was to move the experience from “data reporting” to “decision supporting.” The dashboard materials should answer whether purchased licenses are being used, how adoption is trending, where engagement is strong or weak, which roles or specialties show stronger adoption, and whether users are consistently returning.
01
Start with admin decisions, not charts
The foundation of the work was to treat analytics as a decision-making experience. In the dashboard guidelines, I framed dashboards as being for a specific audience trying to make specific decisions and emphasized starting with the user questions the dashboard must answer before selecting visuals.
For Dragon Copilot, that meant shaping analytics around engagement and who do we need to support? Is adoption growing or declining? When are things going well and when do we need to intervene? These questions align with the physician dashboard’s stated capabilities around license health, actionable engagement insights, adoption visibility, user-level drilldowns, and role/specialty insights.
02
prioritize high-value signals over exhaustive reporting
Initial research was conducted with potential users where it was discovered that analytics should not overwhelm admins with every measurable detail. It should surface the signals that help them understand health, direction, and action. This matches the DAC dashboard guidance, which says dashboards should surface a small set of decision-grade metrics rather than everything measurable. I decided that session-level and deep workflow analytics should be deprioritized in favor of surfacing high-value signals first, with plans to evolve the experience over time.
03
Make engagement understandable and actionable
The Dragon Copilot analytics materials categorize users into High, Moderate, Low, and No Engagement based on usage volume, frequency, and consistency, with the intent of signaling next best actions such as reinforcing adoption, targeted onboarding, or immediate activation.
From a UX perspective, this was important because it turned abstract usage data into a more understandable model for admins. Rather than asking admins to interpret raw sessions or activity counts, the experience could help them see where adoption was healthy and where support was needed. The visualizations should tell a story, not just show numbers.
04
Design for license utilization and adoption health
The dashboard included license health at a glance, including total versus assigned licenses, active users, license utilization trends, and gaps such as unassigned licenses, users who never started, drop-risk users, and users with multiple licenses.
This created a clearer bridge between product usage, operational adoption, and business value. It helped admins understand not only whether people were using Dragon Copilot, but whether their organization was getting value from the licenses they had already purchased or assigned.
05
align the experience with the platform
Dragon Copilot analytics needed to live inside Dragon Admin Center, so the experience needed to feel like part of the platform. The DAC dashboard guidance established rules for report headers, filters, top-down layout, and avoiding redundant left-side navigation inside reports because the DAC shell already provides navigation and page context.
One challenge was balancing simplicity and depth. Customer admins needed a quick overview of adoption and license health, but research also showed a need for deeper segmentation, customization, and feature-level visibility. Customer interviews identified issues such as inflexible segmentation, shallow usage visibility, limited AI quality visibility, reactive system health, and manual bridging between insight and action.
Another challenge was designing analytics for multiple personas and licensing contexts. Users need a shared, simplified analytics report intended to serve both nurses and physicians, with value tied to consistency across nursing and physician Dragon Copilot users.
There was also a platform challenge. Dragon Copilot analytics needed to fit within DAC patterns while still serving product-specific needs. The physician overview report should provide a standardized, unified analytics experience across license types and be accessible through centralized UI navigation in Dragon Admin Center.
final solution
After a few iterations, I landed on a quick and scannable report for the first release of the Dragon Copilot for Physicians analytics.
At the top of the page, high-level KPI cards provide an immediate snapshot of adoption and license health. These metrics answer the most common questions administrators need to understand within seconds of opening the dashboard, such as whether licenses are being used, how many users are active, and whether adoption is increasing or declining. The goal was to make the overall health of the organization visible at a glance without requiring users to interpret charts or drill into details.
Below the KPI section, the dashboard shifts from what is happening to why it is happening. Trend visualizations and breakdowns help administrators explore adoption and engagement patterns across time, roles, specialties, and user segments. Rather than serving as passive reporting, these visualizations were designed to surface opportunities for intervention, helping organizations identify where adoption is strong, where engagement is declining, and which groups may benefit from additional support.
The lowest section of the experience supports investigation and follow-up actions. Detailed tables provide visibility into individual users and licenses, allowing administrators to move from organizational trends to specific people who may require onboarding, re-engagement, or license reassignment. By organizing information in this hierarchy, the experience supports both leadership-level monitoring and day-to-day operational decision making within a single reporting surface.
consolidate first, then deepen
The physician dashboard replaced separate usage reports with a single consolidated view intended to help admins quickly identify adoption trends, engagement gaps, license optimization opportunities, and where action is needed.
use engagement tiers to translate behavior into action
The experience grouped users into engagement levels designed to indicate next best actions instead of expecting admins to interpret raw usage data alone.
keep the top of the dashboard focused on fast comprehension
The page should be formatted in a pyramid, where the top of the page is the quick and scannable information (KPIs, quick signals, etc.) and the bottom of the page should be the deeper and more granular information (tables, spreadsheets, etc.).
support user-level drilldowns where action requires specificity
The Dragon Copilot analytics materials needed filtering and exporting user-level data so admins can identify users or groups that may need support.
outcomes & reflection
Dragon Copilot customer-facing analytics went live in Dragon Admin Center in the US, Canada, and UK, enabling customers to self-serve insights needed to drive implementation and adoption of Dragon Copilot. The work established a path for embedded Power BI reports in the Microsoft tenant on Dragon Admin Center, creating a smoother path for other product and engineering teams to follow the pattern.
The physician-specific Dragon Copilot Overview Dashboard was later described as a purpose-built analytics experience in Dragon Admin Center that brought adoption, engagement, and license health into a single actionable view for admins. From this, Dragon Copilot for Nursing and Dragon Copilot for Radiology quickly adopted the same format and principles for their dashboards.
Next project