Building better, together

AI accelerates idea development but risks fragmenting context, decisions, and reviews when speed outpaces coordination. The guide introduces five actionable moves to keep work cohesive, each paired with a takeaway, a practical exercise, and a real-world example. The approach avoids overhauling processes, instead focusing on the next task at hand to maintain clarity and continuity. Teams can apply these moves immediately without waiting for broader adoption or new tools.

One move advises giving AI the correct design context to avoid plausible but misaligned outputs. Using a team’s design system—including hierarchy, interaction patterns, and accessibility standards—ensures generated work reflects the product’s actual language. For instance, a designer prompting an AI prototyping tool with a design system library can produce a first-pass prototype that aligns with existing components and flags gaps, rather than inventing new patterns.

Another move emphasizes keeping context linked to the work rather than scattered across repositories. An AI assistant connected to live briefs, research, prototypes, and code can surface the latest decisions and unresolved questions for the next contributor. This prevents redundant rebuilding of context, such as a product manager, designer, and engineer each recreating the same background for onboarding. The goal is to make the work’s origins and current state transparent, with clear links to prior decisions and open questions.

The final move advocates for automating routine tasks while reserving human judgment for critical moments. AI can summarize prototypes and highlight assumptions, but teams should focus meetings on debating trade-offs and edge cases rather than status updates. For example, an AI review skill can prepare a session by summarizing the approach and open questions, ensuring time is spent on decisions that require human input rather than passive reporting.