LLM App Builder for Novice Users
For this Georgia Tech research project, we designed a no-code LLM application builder that lowers the barrier to AI creation for non-technical users. We focused on usability, accessibility, and integrated learning tools to support novices in understanding, customizing, and deploying their own LLM-powered applications.
Timeline
Sep - Dec 2023 (4 months)
My Role
Product Designer (UX research, UX design, evaluation)
Team
3 Designers (me, Benedicte Knudson, Hang Wang), 2 researchers (Ashley Frith, Katie McIntyre)
Overview
Today’s large language models (LLMs) are powerful, but building custom LLM applications still requires coding knowledge, familiarity with AI concepts, and comfort navigating complex tooling. Many novice users want to personalize or build LLM-driven tools, but feel locked out due to technical barriers.
Our project asked:
How might we support less-experienced users in crafting custom LLM applications using their own data?
My role: as part of a five-person team, I conducted interviews, moderated sketch and wireframe evaluations, and co-led usability testing and cognitive walkthroughs. I synthesized insights into design implications, contributed to the IA, and led the visual design.
