Designing in Ambiguity

This project marked the pilot of an AI Home Search Assistant, built at a time when universal design guidelines for generative systems were still emerging. It offered a rare/valuable opportunity to design an assistant from the ground up, shaping interaction patterns, conversational dynamics, and defining the role AI would play within a traditionally structured search experience.

Working closely with Engineering, I collaborated on prompt experimentation, gaining firsthand exposure to how prompt engineering decisions directly influenced output quality, user trust, and system behavior. This proximity to implementation deepened my understanding of how language, logic, and design intersect in AI products.

I also contributed to the creation of the new conversational design library within the design system, establishing reusable interaction patterns, tone principles, and structural components to ensure consistency and scalability across the AI experience. This foundation enabled the assistant to evolve without losing clarity or coherence.

As a strategic pilot, the project required navigating uncertainty, shifting definitions, and high stakeholder visibility. It meant designing in ambiguity, aligning vision, technical feasibility, and business expectations while continuously learning through experimentation, it was clearly the most challenging part of the project.

More than building an assistant, this initiative strengthened my ability to create structure in emerging spaces and collaborate strategy across disciplines in moments of rapid technological change.

/ This project is under NDA. I'd be happy to share the full story during an interview!