Case study · 02

Tubi - Rabbit AI

Tubi - Rabbit AI

Tubi - Rabbit AI

This overview is intentionally designed for a quick read in under 3 minutes. For the full story, please review the attached case study deck.

Full Case Study

Rabbit AI was Tubi's first GPT-4 powered content discovery experience, designed to help viewers discover content using natural language instead of traditional keyword search. As Lead Product Designer, I partnered with Product, Research, and Engineering to explore how conversational AI could reshape the discovery experience.

Rabbit AI was Tubi's first GPT-4 powered content discovery experience, designed to help viewers discover content using natural language instead of traditional keyword search. As Lead Product Designer, I partnered with Product, Research, and Engineering to explore how conversational AI could reshape the discovery experience.

The Opportunity

Traditional search worked well when users knew exactly what they wanted, but discovering niche content remained difficult. We saw an opportunity to make exploration more intuitive while helping viewers uncover more of Tubi's long-tail catalog beyond the most popular genres.

Rethinking Search

Rather than asking users to think in keywords, we explored a new search model where people could simply describe what they were looking for in everyday language. This shifted search from matching terms to understanding intent.

Designing Rabbit AI

Over eight weeks, I iterated on multiple interaction models to make conversational search feel approachable. We quickly moved away from voice-first interactions, simplified the interface, and focused on helping users discover content instead of learning how to use AI.

Launch Experience

The final experience introduced Rabbit AI as a dedicated destination where users could ask conversational questions, refine recommendations, and immediately jump into playback. Every interaction was designed to keep content discovery fast and seamless.

What We Learned

The launch generated excitement and strong press coverage, but user behavior told a different story. Funnel analysis and research revealed that most viewers preferred familiar search patterns, making the standalone chatbot harder to adopt than we had anticipated.

The Pivot

Instead of treating AI as a separate destination, we reimagined it as an enhancement to the existing search experience. By embedding AI into familiar workflows, we reduced friction while preserving the benefits of conversational discovery.

Results & Reflection

The redesigned experience improved search refinement by 18%, click-through rate by 12%, and unexpected content discovery by 15%. More importantly, the project reinforced an important product lesson: the most successful AI experiences don't replace existing behaviors—they make them feel smarter and more natural.

Explore more selected work.

Back to work

If you think I'd be a great fit for your team, I'd love to connect.
Email me: agenticram@gmail.com

© 2026 Ram

If you think I'd be a great fit for your team, I'd love to connect.
Email me: agenticram@gmail.com

© 2026 Ram

If you think I'd be a great fit for your team, I'd love to connect.
Email me: agenticram@gmail.com

© 2026 Ram