Medium: Designing for 2026 — process, agency, and decision quality in the age of AI

Lee Munroe, VP of Design at OneSignal, published this piece in January 2026 as a practical framework for designers navigating the current shift. He draws on Jenny Wen’s talk at Anthropic and his own experience leading design teams through a year in which prototyping timelines collapsed.

The central argument is that process exists to support learning and decision-making, not to follow steps for their own sake. When AI tools like v0, Figma Make, Replit, Lovable, and Claude Code can generate interactive prototypes in minutes, skipping low-fidelity wireframes makes sense. But skipping the thinking those wireframes were meant to prompt does not. “Speed without clarity leads to thrash,” Munroe writes.

Three ideas carry the article.

The first is that designer agency has expanded significantly. Designers can now explore and validate ideas earlier than any previous generation of tools allowed, which means they also carry more responsibility for the direction those explorations take. Consequential product decisions arrive earlier in the process, not later.

The second is decision quality as the key metric. Faster tools don’t improve outcomes unless they’re paired with clearer thinking. Munroe argues that skipping ceremony is fine, but skipping rigor is not — teams still need to understand what they’re deciding and why.

The third is craft as the differentiator. Since AI can generate adequate UI immediately, “good enough” no longer separates products. Human judgment about taste, coherence, and clarity is what distinguishes work that holds up from work that doesn’t.

The article also cautions against chasing every automation opportunity. Selective adoption matters more than maximum automation. Not every step that can be delegated to AI should be — especially in stages where collaboration with engineers, PMs, or users generates insights that no prototype can surface alone.

Useful for senior designers, design managers, and design leads who are managing teams through rapid AI adoption and want a framework that goes beyond tool lists.