Short Lessons from Design Work in Practice

Field Notes is a growing collection from Design Outcomes. It shares practical ideas from design teams and individuals. New entries appear every Friday. Readers can search 113 entries or browse topics such as research, systems, prototyping, critique, craft, teamwork, storytelling, and hiring. Each note gives a short explanation, key details, and a link to the original article.

Several recent notes examine how teams use AI without giving it full control. Shopify used LLMs to rebuild a 300-screen React Native app in small checkpoints. The process checked behavior, reviewed visuals, tested code, and required engineer approval. Another note says agents need clear rules about what they can do, what they should do, and what they may decide. This makes design-system decisions easier for both people and agents to understand.

Other notes focus on human needs. A Nubank accessibility file guides Claude to write for blind and low-vision users. Responses should make sense when a screen reader reads them from top to bottom. Tables become prose, unclear location words are replaced, and long logs are reduced. A Microsoft Design note presents automation as a way to remove mechanical work and create space, rather than adding more features.

Leadership and judgment are also common themes. One article asks designers to own business outcomes instead of only pleasing stakeholders. Another says context libraries can matter more than perfect prompts for AI agents. Global, local, and changing information need different care. A note about generative UI warns that a screen created in 45 seconds can need three hours of cleanup. Fast output can hide wrong patterns, missing rules, or accessibility problems.

Field Notes does not replace the original articles. It helps readers decide what deserves more attention and explains the practical lesson first. Recent topics also include AI safety, human responsibility for edited work, and the value of writing practice. Together, the entries show that good outcomes come from clear decisions, useful constraints, careful review, and learning from real work.