A professional services firm generates thousands of research documents and thought leadership reports — but staff had no efficient way to access them. The task was to design and build a proof-of-concept AI knowledge assistant for the firm's intranet: a conversational tool that let people query the entire document library in plain language.

As Creative Director and Strategist, the design was built around four principles: intuitive (no blank-canvas anxiety, no steep onboarding curve), resourceful (responses that were legible and actionable, not just accurate), accessible (serving everyone from a senior partner with fifteen minutes to an analyst building a deep sector report), and familiar (fitting naturally within the firm's existing digital ecosystem). Key design decisions included a dual-layer citation system that grounded every response in verifiable sources; AI-generated follow-up prompts to encourage deeper exploration; a 28-day conversation window to prevent stale outputs being treated as current; and a staged loading animation that made the AI's processing time feel purposeful rather than broken.

The result turned the firm's static document library into a dynamic, conversational research resource.