[Case 02]
From Fragmented AI Assets to a Unified Discovery Experience
Consulting / AI Technology
aIQ Hive Revamp
Revamping KPMG’s AI App Catalog with Smarter Search, Personalization, and Community Features
[Project Overview]
aIQ Hive is KPMG’s centralized portal for discovering, learning, and using the firm’s AI apps, assets, agents, and data. In the latest revamp, I focused on enhancing the user experience by making content easier to find, increasing engagement, and delivering personalized recommendations. Leading the design effort, I introduced an AI-powered universal search, asset-level AI Q&A, community threads, tailored suggestions, and a feedback system. The result is a streamlined, responsive platform that makes navigating KPMG’s AI ecosystem faster and more intuitive.
[Problem Statement]
As KPMG’s portfolio of AI tools, assets, and assistants rapidly expanded, users found it increasingly challenging to locate relevant solutions quickly. The existing aIQ Hive lacked personalized recommendations, community engagement opportunities, and a simplified way to search across the catalog. Users wanted a more intuitive, intelligent search experience and a space to collaborate and share insights. The goal was to redesign the Hive to address these gaps—enabling faster search, providing tailored asset suggestions, fostering community interaction, and creating feedback loops to inform continuous improvement.
[Industry]
Consulting / AI Technology
[My Role]
Lead Designer
[Platforms]
Desktop, Mobile & Tablet
[Timeline]
December 2024- March 2025
[Persona]
Julia Jones
Engagement Manager
I need a fast, intelligent way to find the right AI tools for my project—without having to dig through lists or outdated pages.
Age: 41
Location: Philadelphia, PA
Tech Proficiency: High
Gender: Female
[Goals]
Quickly find relevant AI tools, assets, and assistants tailored to project needs.
Get detailed, context-specific answers about each AI asset without leaving the page.
Engage with peers, share experiences, and provide feedback to enhance the AI Hive ecosystem.
[Frustrations]
Difficulties locating specific assets among an overwhelming catalog.
Limited guidance or tailored recommendations based on personal use or history.
Lack of peer insights or community-driven knowledge sharing around AI assets.
[Process]
[01] User Research
Conducted interviews with AI practitioners, engagement managers, and client professionals to understand common challenges when navigating the aIQ Hive.
Analyzed behavioral data and search patterns to pinpoint drop-off points and high-friction areas in the user journey.
Facilitated co-design workshops with internal stakeholders to gather ideas for improving search functionality, personalization, and community engagement.
[02] Insights
Users needed faster, more intuitive search experiences that surfaced both AI-generated summaries and related assets.
There was a strong desire for personalized recommendations to support discovery of new AI assets relevant to their roles and previous activity.
Community features and direct feedback mechanisms were essential to foster user engagement and collect insights to improve the platform.
[03 Design Solution]
Embedded AI-powered search with suggested keywords, recent activity, and dynamic AI-generated responses modeled after Google’s AI Overview.
Added asset-level embedded Q&A search (inspired by Amazon’s Rufus) to help users ask targeted questions and receive asset-specific information.
Launched Community Threads (Reddit-style) and Feedback features to enable knowledge sharing, real-time feedback, and continuous engagement.
[04] Testing & Iteration
Conducted usability testing with advisory professionals and AI users to validate new search flows and community features.
Ran A/B tests on AI search result placements to determine optimal visibility and user engagement.
Collected ongoing feedback via in-tool surveys and the new feedback feature, continuously refining the user experience based on user input.
[Outcomes]
Reduction in time spent searching for AI assets, as reported by post-launch user feedback.
Increase in user engagement, with more interactions in Community Threads and higher adoption of suggested asset recommendations.
Established a scalable framework for AI-powered search and personalization, setting the foundation for future enhancements and integrations.
[Key Learnings]