NotebookLM Design Case Study by Jason Spielman Jason Spielman led the design of NotebookLM, shaping its core user experience, brand identity, and visual system from experimentation through launch in 2024. NotebookLM is recognized as one of TIME’s Best Inventions of 2024. --- Overview Role: Design Lead - UX and Identity Year: 2024 Designed a completely new product from the ground up, exploring fresh paradigms and inventing new interaction patterns. Collaboration with a tight-knit, cross-functional team was key. --- Problem + Requirements Addressed "tab overwhelm": the fragmented, fractured experience users have while jumping between multiple tools for synthesis and creation. Goal: Build an integrated space allowing reading, writing, and creation in one tool, to cover the entire user journey. Inputs: Source list Open sources (wide view) Outputs: Notes list Open note (wide view) Core features: Chat interface with citations Multiple entry points for creation workflows --- Mental Model The product is structured around the creation journey: Inputs: User brings in sources like documents, notes, and references. Chat: Users engage with their sources through conversational AI (ask, clarify, synthesize). Outputs: Generate structured outputs like notes, study guides, and Audio Overviews. This linear yet flexible flow (Inputs → Chat → Outputs) provides users with a clear mental model and keeps AI interactions digestible. --- Design Solution: 3 Panel Structure Designed a responsive panel system adapting to user needs. Panels scale fluidly, maintaining quick access to sources and notes even on smaller screens. The 3 key panels: Source Panel: All user-provided sources live here; the start of a user’s journey. Studio Panel: Workspace where inputs transform into outputs. For creating, shaping, and capturing work. Chat Panel: Central to the experience; dynamically resizes based on focus and task. Panel States and Layout Variations Panel layouts dynamically adjust to fit different workflows: Standard: Balanced view of sources, chat, and notes. Reading + Chat: Optimized for citing sources and generating responses. Chat + Writing: For drafting and iteration (a popular user request). Reading + Writing: Ideal for composing while keeping sources visible. The system is designed for scalability, accommodating new tools, modes, and workflows without breaking the structure. --- UI Evolution Early UIs featured exploratory chat overlays and notes-driven designs. Iterations synthesized into the scalable 3-panel layout. The design journey included thousands of iterations and sketches, including solutions sketched on napkins to find clarity. --- Audio Overviews Feature Jason led the design of the Audio Overviews, from prototype to launch. Introduced new interaction paradigms such as "interrupt." Audio Overviews provide a one-click way to study and digest information, grounded in source data. This feature rapidly gained viral attention. The name “Audio Overview” was unintentionally coined by Jason. Shoutouts to the cross-functional team including Stephen Hughes, Oliver King, Biao Wang, Corbin Cunningham, Yi Yao, Usama Bin Shafqat, Simon Tokumine, Michael Chen, and others. --- Brand Identity Developed quickly in collaboration with Google Labs and the Google central brand team. Key contributors: Feel Hwang, Nick Mcginnis, Jennifer Leartanasan, and team. Created full suite of visual assets, from press kits to launch visuals and animations. --- Key Takeaways from the Journey Building with users, not just for users Ship early, iterate fast. Anticipate user needs like in-line citations quickly. Engage actively with user feedback and community (e.g., Discord). Built-in, not bolted-on AI is