M365 Agent Store
Designing Discovery and Acquisition for AI Agents in M365
About the project
Company
Microsoft
Timeline
2024 - 2025
My role
UX designer, AI designer
Challenges
Context
As the number of AI agents grew, users faced choice overload and uncertainty about which agent fit their needs. The store needed to feel curated and contextual
This presented 3 key design challenges:
Challenge 1
How to balance browse vs. recommendation?
Challenge 2
How to build trust signals for unfamiliar AI agents?
Challenge 3
How to make discovery feel integrated, not disruptive?
The designs
Due to confidentiality, I can only share a limited view of the project. Here are a few examples of the flows and designs our team created.
I’d be happy to walk you through the details and answer any questions during a call.
Agent Store
The experience makes discovering AI agents intentional, trustworthy, and effortless. It helps users find the right agent at the right time without overwhelming them with a catalog.
What I did
Created all-new Agent Store, entry points and added trust signals.
Created a single-column layout optimized for quick scanning, with clear hierarchy for agent details.
Simplified navigation, introduced filters, search and designed different states with helpful suggestions.
Impact
Reduced “catalog fatigue”.
Shifted discovery from list-heavy browsing to contextual, confidence-building experiences.
Increased user trust through transparent provenance and clear agent capabilities.
Before vs after
Prior to our team's effort, Agents lived in the Teams app store, which created confusion and an overwhelming experience for users.
Teams app store
M365 Agent store
Confidence-building architecture
By structuring content into clear tiers, we created a sense of clarity and predictability. Users can start with familiar agents, explore curated options that signal quality, and discover personalized recommendations that feel relevant to their work.
Quick access to most used agents
Featured content
Tailored recommendations
Results
Metrics
Engagement
Increased click-through on in-chat recommendations.
Increased store visits from contextual entry points.
Conversion
Reduced drop-off from browsing to activation.
Increased activation rate for recommended agents.
Efficiency
Less time-to-find relevant and useful agents.
Increased search success rate after filter redesign.
Next steps
Personalized Discovery
Use role-based and activity-based signals for smarter recommendations.
Improved search
Enable natural language queries for finding agents.
Cross-Surface Consitency
Extend Store patterns to Teams, Outlook, and other M365 canvases for unified discovery.
Key takeaways
Context beats catalogs
Users prefer timely, explainable suggestions over endless lists and tailored useful recommendations over over-the-counter
Discovery is powered by trust
Users ignore catalogs; they respond to timely, explainable suggestions. Context beats quantity.
Trust is visual
Authorship and usage cues matter more than clever copy. Users will trust context-based recommendations.
Discovery is a system
It spans chat, store, search and more, not a single screen.
