Copilot First-Run Experience
As part of the Copilot design system team, I contributed to defining the first-run experience patterns—establishing reusable components and interaction guidelines that ensure consistency across all Copilot touchpoints in Azure.
Identified Needs
Through user research and stakeholder interviews, we identified three core needs for the first-run experience. Users needed a clear understanding of what Copilot can do, an easy way to start interacting, and opportunities to discover more capabilities over time.
Simple Introduction
Provide a brief introduction of Azure Copilot to help users understand its capabilities.
Prompt Starters
Provide prompt starters for users to easily submit a request and get started.
Prompt Discovery
Enable opportunities for users to explore additional prompts and capabilities.
Design Solutions
Global Entry Point
Discoverability was a key challenge—users couldn't engage with Copilot if they didn't know it existed. We designed a teaching notification that appears when first-time users log in, positioned near the Copilot button in the portal header. This creates awareness at the right moment without disrupting the user's workflow.
Walkthrough Dialog
When users first open Copilot, a three-step dialog introduces key capabilities in digestible chunks. Each step highlights a different use case—asking questions, getting recommendations, and automating tasks. The walkthrough concludes with a clear "Try Copilot" call-to-action that disappears after the user's first interaction, respecting their time on subsequent visits.
Adaptive Prompt Starters
Blank text fields can be intimidating—users often don't know what to ask. We designed prompt starter cards that evolve with the user. First-time users see curated static suggestions covering common tasks, while returning users receive personalized recommendations based on their Azure resources, recent activity, and trending requests across the platform.
Contextual Assistance
Context is everything. When users access Copilot through in-context nudges placed throughout the portal, the chat opens with prompt suggestions tailored to their current task. For example, if a user is viewing a virtual machine, they'll see prompts related to VM management, monitoring, and optimization—reducing cognitive load and demonstrating immediate relevance.
