Xnode: An AI-assisted Collaboration Platform for XR Production Teams

Key Visuals

Background / Problem

XR content production involves multiple roles with different domain vocabularies and fragmented context across meetings and artifacts. The team identified recurring bottlenecks in cross-functional communication, shared terminology alignment, and decision traceability during fast-paced project cycles.

My Role

I contributed as a user researcher in UXR Lab, helping define research questions, synthesize collaboration pain points, and shape the user-centered requirements for the platform's assistant workflows.

Process / Method

The project followed a research-to-product flow including observation study, desk research, focus group interviews, affinity diagram sessions, in-depth interviews, prompt analysis, prototype testing, and MVP refinement. Findings were translated into product goals: domain-aware response generation, team language dictionary building via labeling, and context timeline support using meeting minutes and project data.

Result / Outcome

The team proposed an architecture combining preprocessing pipelines, RAG retrieval, vector search (Supabase), and assistant orchestration (Claude API/Gemini + file search) to support project Q&A, meeting summarization, and context-aware collaboration. A demo and staged roadmap were defined for real deployment in XREAL workflows.