Designing for AI search surfaces like Google's AI Mode requires abstracting UI into modular, portable primitives. Rather than designing isolated containers for specific data types like web results, designers must conceptualize flexible interaction systems capable of ingesting diverse ecosystem content. This prevents component fragmentation and ensures consistency across both core search results and emerging AI surfaces.

Power users of AI agents believe the ideal user interface is not graphical but conversational. They prefer text-based interactions within existing chat apps and see voice as the ultimate endgame. The goal is an invisible assistant that operates autonomously and only prompts for input when absolutely necessary, making traditional UIs feel like friction.

While designers explore novel interfaces, John Bai's experience building GrokBot reaffirmed the power of the traditional multi-column chat UI. He concluded it's still the "right way" to interact with a fleet of AI agents carrying out distinct, predefined tasks.

The belief that chat is the ultimate UI is a projection from high-agency builders like Sam Altman and Elon Musk. Most consumers aren't looking to save time but to spend it. They prefer browse-based interfaces for discovery and entertainment, not command-line efficiency, which represents a major builder bias.

To appeal to the "layperson" rather than tech early adopters, Comet's designers made the core browser experience familiar, like Google Chrome. This reduces cognitive load, allowing users to focus their limited learning bandwidth on the novel AI features, even if it disappoints power users expecting a radical redesign.

While AI enables fully dynamic, non-deterministic interfaces, this isn't always desirable. Core workflows like login, billing, and settings require stability and predictability. A key design skill is now discerning what should be a fixed, reliable UI versus an adaptive, personalized one.

At OpenAI, the first question is "Can we solve this with the model (tokens) instead of pixels?" This treats the AI as the primary design material, pushing designers to think about interaction and behavior before creating bespoke user interfaces.

Giving AI freedom to generate custom layouts dynamically risks disorienting users if interface patterns shift with every response. Rhiannon Bell explains that while AI can assemble components from design systems, designers must enforce human-defined rules. Predictability in where answers, interactive widgets, and links appear is critical so users do not have to learn a brand-new interface on each query.

AI models are poor at "last-mile" visual design. However, if a human designer invests heavily in creating a perfect set of primitives (e.g., buttons, cards), AI becomes incredibly effective at reusing and intelligently extrapolating from that foundation for new contexts. Human effort on the core system pays off exponentially.