Jun 26, 2026Economic ResearchAnthropic Economic Index report: Cadences

2026-06-25 · Source: Anthropic Research · Field: Finance & Economics — Economic Analysis & Policy, Artificial Intelligence & Machine Learning, Data Science & Analytics · Depth: Expert, extended

Summary

The Anthropic Economic Index report, "Cadences," published June 26, 2026, details evolving Claude usage patterns and economic impacts, adapting its data pipeline to capture long-running agentic tasks. The report introduces hourly sampling, a new output classifier, and granular data for chat, Cowork, and 1P API, alongside initial findings from an April 2026 user survey. Key findings reveal Claude usage mirrors daily and weekly rhythms, with personal queries spiking on weekends and work-related tasks in higher-wage occupations increasing during non-traditional hours. The study classifies 93% of conversations as producing artifacts like explanations (17%) or documents (15%), noting that compute consumption scales with the estimated value of work. Claude Code exhibits higher AI autonomy (0.37 points average difference) compared to chat/Cowork, and Claude's responses generally show a one-year higher reading level than user prompts. Survey results from ~9,700 users indicate over 35% anticipate AI handling "most" of their work within 12 months, with those delegating more tasks to Claude expressing greater optimism about future job outcomes.

Key takeaway

For AI Product Managers developing agentic capabilities, this report suggests that enabling higher AI autonomy in complex, high-value tasks can significantly boost user optimism regarding job security and pay. Focus your development efforts on features that allow Claude to handle more intricate workflows, as this correlates with users feeling their skills are more valuable and learning at the same rate. Prioritize tools like Claude Code that facilitate deeper delegation.

Key insights

AI usage patterns reflect economic cadences and task complexity, with higher delegation correlating to user optimism.

Principles

• AI usage mirrors real-world work and life rhythms.
• Compute cost correlates with task value.
• Increased AI autonomy can foster user optimism.

Method

Anthropic updated its data pipeline to sample Claude usage hourly, classify conversation outputs into over 30 artifact types, and link usage data with survey responses using privacy-preserving methods.

In practice

• Analyze AI usage logs for daily/weekly work patterns.
• Classify AI outputs to understand value creation.
• Explore AI delegation in high-value tasks.

Topics

• Anthropic Economic Index
• Claude Usage Patterns
• AI Autonomy
• Workforce Impact
• AI Artifacts
• User Perception

Best for: Research Scientist, Director of AI/ML, AI Product Manager

Related on AIssential

• Mar 31, 2026Economic ResearchHow Australia Uses Claude: Findings from the Anthropic Economic Index — Australia exhibits high per capita Claude AI adoption, driven by diverse, complex tasks in professional and personal use, with less delegation.
• Mar 24, 2026Economic ResearchAnthropic Economic Index report: Learning curves — Experienced AI users achieve greater success and tackle more complex tasks, indicating a learning curve in AI proficiency.
• Anthropic’s Claude is winning over paid consumers, a market owned by ChatGPT — Anthropic's Claude is rapidly gaining paid consumer market share and mindshare, challenging ChatGPT's dominance.
• Claude Code Killed the AI Bubble — AI agents, particularly Claude Code, are transforming software development and broader information work, signaling a major economic shift.
• Anthropic Raises $30 Billion - What It Means for Indian IT? — AI companies like Anthropic are disrupting traditional IT services by scaling output through intelligence, not headcount.
• Experienced Claude users better at harnessing AI tool, says Anthropic - Business Standard — Experienced users develop strategies to better harness AI tools like Claude, leading to improved outcomes.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Anthropic Research.