A conversational AI experience that modernizes how fans discover games, players, and history, giving Major League Baseball fans instant, personalized access to live data, archives, and predictions across mobile and web.
Launch Year
2024
Platform
Mobile, web
Role
Senior Product Designer
Problem
MLB was introducing conversational AI for the first time in a space defined by deep tradition, dense data, and established fan behaviors. The AI Chat Assistant needed to transform decades of stats and media into an intuitive, trustworthy conversational experience—one that felt modern and exploratory while preserving the accuracy, familiarity, and credibility fans expect from MLB.
Outcome
The MLB AI Chat Assistant introduced a conversational entry point into baseball fandom, allowing fans to explore live games, historical stats, player stories, and predictions through natural language instead of dense tables or static pages.
This marked a shift for Major League Baseball: appealing to younger fans who expect interactive, AI-driven products while preserving the trust, accuracy, and familiar UX patterns longtime fans rely on. The assistant needed to feel modern without compromising data credibility or brand heritage.
The core challenge was designing a conversational experience that balanced speed and exploration with structure and reliability, enabling fans to move fluidly between live information, history, and hypothetical scenarios without losing context.
My Role
I was the Lead UX/Product Designer on the MLB AI Chat Assistant, working alongside one additional UX designer and a creative director over an eight-month engagement.
I owned the experience end to end, from early product definition through interaction design, modular UI systems, and usability validation. Because the assistant relied on real-time data, historical archives, and predictive content, my role extended beyond interface design into defining how AI responses should surface, structure, and transition between conversational and visual states.
A critical part of my role was ensuring conversational flexibility did not undermine clarity, accessibility, or trust—especially when adapting established MLB UX patterns into a chat-first paradigm.
My Design Focus
Making conversational discovery feel trustworthy
This product could not succeed as a novelty chatbot. Research showed that fans were open to conversational exploration, but only if answers felt authoritative and grounded in MLB’s data ecosystem.
I focused on pairing conversational responses with structured UI modules that reinforced credibility—clearly communicating what was live data, historical context, or predictive insight. By anchoring AI responses in recognizable MLB formats, the assistant felt additive rather than disruptive to existing fan workflows.
Designing flexible paths without losing context
Fan behavior was highly nonlinear. A single session often jumped from game times to player stories, probable pitchers, ticket purchasing, and back to live scores.
I designed the experience to support branching by default, allowing fans to explore multiple directions without dead ends while preserving conversational history and continuity. This made exploration feel fluid rather than transactional and encouraged longer, more engaged sessions.
Scaling personalization across teams, players, and eras
One of the most complex problems was designing for scale: 30 teams, decades of history, and over 20,000 active and retired players.
Through story mapping and iterative testing, we moved from rigid templates to a flexible system that dynamically adapts team branding, stats, and layout based on player type and historical relevance. This allowed each interaction to feel tailored while remaining consistent, performant, and maintainable.
Designing iteration into the core experience
Fans rarely ask one question and leave. Exploration naturally unfolds through follow-ups.
Rather than treating follow-up questions as edge cases, I designed iteration as the default loop—allowing users to move seamlessly from high-level questions into deeper detail without restarting or recontextualizing. This reinforced the assistant as a discovery tool rather than a simple Q&A interface.
How this work affects my design approach
I was the UX Designer responsible for Create with Alexa, covering generative writing and image creation across Echo devices, the Alexa app, and the web. I owned how customers initiate creation, refine outputs through conversation, and build confidence in what Alexa produces.
Because this was Alexa’s first generative AI product, output quality and UX were tightly linked. A key part of my role was partnering closely with engineering and applied science to shape prompt instructions and refinement behaviors so writing outputs were clearer, more intentional, and easier to correct when they missed the mark. While I did not train models directly, my work focused on ensuring the experience reliably produced competitive results.