Taco Bell
From static menu boards to an AI-powered dynamic drive-through experience — and how design made a $1B+ opportunity legible to the people who needed to believe in it.
The drive-through menu board is one of the most high-stakes pieces of real estate in quick service retail — and for most of the industry, it's been treated as a static poster that happens to be digital. The Taco Bell engagement I led across two phases set out to change that entirely: first by establishing the right design foundation and strategic framework, then by building and deploying what the team positioned as industry-first AI-driven dynamic menu capabilities, with a projected $1B+ revenue opportunity over five years.
Taco Bell's existing boards had a fundamental problem: the more you show customers, the less they see. I established four strategic guideposts — Entice Customers to Discover, Right Size My Meal, Simplify Everything, and Give Customers Control — and built a modular, atomic design-based component system to support the dynamic capabilities Phase 2 would require. Rather than designing a single fixed layout, I presented solutions on a spectrum, allowing stakeholders to pressure-test tradeoffs rather than simply react to a recommendation.
My role in Phase 2 shifted from execution to translation. The engineering and data science teams were building an AI-driven decision engine capable of serving real-time, store-specific menu recommendations at scale. The challenge wasn't just technical — it was organizational: getting Taco Bell's C-suite, franchisees, and internal teams to understand what this technology could actually mean for them.
I led the development of stakeholder vignettes — narrative experience scenarios told from the perspective of the people this system would affect most: an unknown customer becoming a known one through progressive personalization, a franchisee using real-time data to respond to a staff shortage, a merchandising team running rapid AI-assisted tests instead of waiting weeks for results.
"These weren't technology demonstrations. They were human stories that made the capability tangible before a line of production code was written."
An in-market pilot was deployed at a live Taco Bell location — a 5-week live test with a 50/50 traffic split, drawing on 5.6M transactions across 95 stores to train the underlying model.
