INSIGHTS
Weekly notes on design and AI.
Ideas for your next decision, drawn from shifts in design and technology.
Every Monday, we look at developments in design and AI and summarize the facts we can confirm, what we read from them, and actions you can try in your own work.
Deciding what to hand to AI by the effort of checking and the criteria for judging
We look at a study of UX practitioners, a method for refining evaluation criteria with concrete examples, Figma's prototype-to-code connection, and the availability terms for an enterprise agent, and sort out what to delegate and how to verify it, starting from the checking and correction work that AI does not necessarily make lighter.
Turning tools built with AI into something a team can keep using
We look at Figma's internal publishing and version history, the general availability of GitHub Agentic Workflows, usage analytics in Claude Enterprise, and a company case study on managing agents, and sort out how to design sharing, evaluation, budgets, and retirement so a team can keep using tools built with AI.
Before handing work to AI, align procedures, permissions, and the scope of verification
We look at the transfer of Agent Skills, a shared format for work procedures, Zoom's plan to connect conversations with follow-up work, Figma's new admin controls, and formal verification done with AI, and sort out the boundaries of procedures, permissions, and evaluation worth settling before automation spreads, along with small ways to test them.
Making work built with AI something the whole team can judge
We look at Figma's new features for sharing code and specs, Mistral's acquisition of a design firm, and Anthropic's approach to third-party evaluation, and sort out the practical decisions and small experiments that turn the speed of building with AI into comparison, agreement, and verification.
From individual time savings to organizational results: rethinking the impact of AI
Faster individual work does not automatically lead to better organizational results. Drawing on five developments from this week, from workflow redesign to quality control in development and design system upkeep, we outline how to check whether AI is really paying off, and what to try next.
Handing work to AI: designing delegation, verification, and stopping
Entrusting AI with ongoing work requires designing not only goals and permissions, but also how progress is understood, verified, and stopped. Drawing on a unified front door for public services, execution environments, talent development, and specialized AI, we consider how much can be delegated.
Building fast with AI: design evaluation and accountability first
AI can produce prototypes quickly, but deciding whether to move into production is a separate step. Drawing on evaluation criteria for PoCs, expert scoring standards, and the roles and incident response for agents, we consider how to turn production speed into operations with quality and clear responsibility.
When business interfaces keep changing, how do you keep the rules for judgment stable?
As business interfaces adapt to conversations and goals, how do you keep permissions and decision criteria intact? Drawing on generative UI, task-specific models, codes of conduct, and incident reporting, we sort out which parts of the display can change, which rules should stay fixed, and how to test them in practice.
Making AI part of how work gets done: connecting data, execution, and deliverables
AI is starting to connect with business data, execution environments, and the formats of finished work. Drawing on agent platforms, enterprise data analysis, industry-specific services, voice and image creation, and the impact on external services, we outline what to design and evaluate when building AI into real work.
Putting AI to work: designing context, permissions, and handoffs
Building AI into everyday work depends on more than model performance. It depends on the context the AI draws on, the permissions it has, and the conditions for handing off to people. Drawing on design system tests, customer service research, and corporate governance policies, we outline the design questions to settle before adoption.