Why it matters
NDC AI field report about giving a development team Claude Code for a day. It is useful because it grounds AI-coding adoption in observed workflow change rather than hype: what gets faster, what review burden appears, what breaks, and how developers adapt their process.
My takeaway: Evaluate coding-agent adoption by watching real teams use it. Productivity claims need to be balanced against review load, context setup, failure recovery, and changes to engineering habits.