India Country Brief: The Anthropic Economic Index
Executive Summary
India represents the second-largest source of Claude.ai usage globally, accounting for 5.8% of total conversations. The country has emerged as one of the world's fastest-growing AI user bases, yet adoption remains geographically and occupationally concentrated. Analysis of ~1 million Claude.ai conversations from November 2025 reveals that Indian users apply AI to substantially more time-consuming tasks compared to global averages, suggesting frontier-level technology utilization.
India's Position in Global AI Adoption
India ranks second globally by total Claude.ai usage volume, trailing only the United States. However, this ranking tells an incomplete story. When adjusted for working-age population, India ranks 101st out of 116 countries with sufficient data. This discrepancy highlights that "India's high overall Claude use reflects population size rather than per-capita adoption rates."
Geographic Concentration
Usage concentrates heavily in economically advanced regions. Maharashtra, Tamil Nadu, Karnataka, and Delhi collectively account for over half of India's Claude.ai use. These states home India's major technology hubs—Bangalore, Hyderabad, Chennai, Mumbai, and Delhi NCR—indicating that "current AI adoption is driven primarily by India's established technology workforce rather than broad-based consumer uptake."
Task and Occupational Patterns
Indian Claude users demonstrate a distinctive occupational profile. Software development tasks dominate, representing 45.2% of all O*NET-mapped activities—the highest proportion globally. Educational and instructional tasks also rank prominently, suggesting learning applications extend beyond professional development.
The concentration in software-related work aligns with India's established IT services sector, pointing to AI adoption following existing professional strengths rather than expanding into new economic domains.
Economic Primitives: How India Uses AI Differently
Analysis reveals several distinctive patterns in how Indian users leverage AI:
Productivity Speedup: Indian users complete tasks in approximately 14.8 minutes that would otherwise require 3.8 hours—a 15x acceleration compared to the global 12x average. This suggests "AI is delivering outsized productivity gains on more complex tasks."
Work Orientation: 51.3% of Indian Claude use relates to professional contexts versus 46% globally. Coursework accounts for 20.9%, with personal use at 27.8%. This work-heavy profile reflects India's professional services sector structure.
AI Autonomy: Indian users grant AI higher decision-making autonomy (3.60 on a 1-5 scale) compared to the global average of 3.38, indicating "greater willingness to let AI operate independently."
Task Difficulty: Only 84.6% of Indian tasks could theoretically be completed by humans alone, versus 87.9% globally. This suggests users frequently bring AI challenges exceeding their independent capabilities.
Prompt Quality: The education level required to understand Indian user prompts (12.2 years) closely matches the sophistication of Claude responses (12.5 years), with India ranking in the top 10% for response sophistication globally.
Key Implications
Broadening Economic Impact: Expanding AI benefits requires moving beyond software and IT services. Current concentration in four states and software-related tasks suggests adoption represents an extension of existing strengths rather than sectoral diversification.
Structural Barriers: The gap between absolute and per-capita usage reflects barriers related to income, digital infrastructure, and awareness outside IT sectors. "Without addressing these structural obstacles, Indian AI adoption is likely to remain concentrated."
Skills Investment: Strong correlations between prompt quality and response sophistication suggest that "training programs focused on effective AI use...could meaningfully improve returns from wider adoption."
Methodology
This analysis draws on privacy-preserving data from Claude.ai consumer usage during November 13–20, 2025, documented in the fourth Anthropic Economic Index report. Geographic assignments rely on IP-based geolocation, while task and occupation classifications map to O*NET taxonomies and SOC occupation groups. Analysis includes Claude.ai Free, Pro, and Max usage.