Help Pane - Microsoft 365
Designed scalable content experiences for M365's Help Pane, conversational systems, and brand channels, bridging user intent with clear language at scale.
Microsoft 365 Help Pane: Transforming In-App Support Through Data-Driven Content Design
Introduction
Microsoft 365's in-app help pane serves millions of users who need quick answers without leaving Word, Excel, or Outlook. As Content Designer and Program Manager, I transformed this experience by authoring 250+ help articles informed by machine learning insights and user behavior data.
Challenge
The help pane wasn't working. Users couldn't find answers to common questions, search results missed the mark, and support tickets piled up for issues that should have been self-solvable. The content existed in isolation from actual user needs, with no clear connection between what people searched for and what we provided.
The brief: Make help actually helpful—and prove it with data.
Solution
I partnered with ML teams to analyze thousands of user queries, identifying patterns in how people actually asked for help versus how we'd written our content. This revealed critical gaps: users searched for tasks, but our content was organized by features.
I rewrote and optimized 250+ articles using real user language from the query data. Titles became action-oriented. Keywords matched intent, not internal terminology. Structure prioritized the fastest path to resolution.
Working cross-functionally with engineering, I helped refine the search model itself—turning query insights into better content retrieval. Every article was informed by what users actually needed, not what we assumed they'd ask.
Conclusion
Results: 30% increase in customer satisfaction, 3% reduction in support costs, and measurably better search accuracy.
Impact: Transformed the help pane from a content repository into an intelligent support tool. Established a sustainable model where user data continuously improves content, and content insights inform product development—closing the loop between what users need and what we deliver.