Anthropic's Economic Index: What 1 Million Claude Conversations Reveal About AI at Work

% of all job categories now use Claude for at least a quarter of tasks, indicating AI became workplace infrastructure, not niche tool. Top 10 tasks dropped from 24% to 19% of conversations, showing AI use diversifying across more applications rather than deepening in same areas. Users with 6+ months experience show 10% higher success rates and use Claude collaboratively as thinking partner versus newer users' one-shot directive approach. Personal use surged from 35% to 42% while coursework dropped, driven by casual queries like sports scores flooding in from growing user base. API automated business workflows doubled while serious programmatic access users select models deliberately based on task complexity, unlike casual Claude.ai users. Anthropic just dropped their March 2026 Economic Index report. It's based on 1 million sampled conversations from Claude.ai and the API, collected between February 5-12, 2026. Three months after Opus 4.5 launched, right as Opus 4.6 went live. Here's w