On the sidelines of the AI Impact Summit 2026 and the opening of the Bengaluru office in India, Anthropic has published its ‘India Country Brief: The Anthropic Economic Index’ report, offering the first detailed picture of how Indians are using its flagship Claude.ai model. The brief draws from anonymised data in the fourth Anthropic Economic Index report, analysing approximately 975,160 global Claude.ai conversations from November 13–20, 2025, including 58,098 from India.
The findings position India as a standout in global AI adoption. India accounts for 5.8% of total Claude.ai consumer use worldwide, ranking second only to the United States. This makes it one of the fastest-growing AI user bases, thereby contributing to the country’s status as the world’s largest exporter of IT services.
However, the per-capita adoption (adjusted for working-age population) tells a different story. India ranks 101st out of 116 countries with sufficient data, far below peers like Singapore or Malaysia. Anthropic attributes the high absolute volume to India’s massive population rather than broad, widespread use among individuals.
Claude used primarily by IT sector in India
Claude’s usage remains heavily concentrated in tech-centric states. It is the state of Maharashtra that leads with a share of 15.5% of India’s Claude interactions, followed by Tamil Nadu (13.2%), Karnataka (12.7%), and Delhi (10.5%). These states account for over half of national activity. The patterns mirror India’s established IT hubs, including Bengaluru, Hyderabad, Chennai, Mumbai, and the Delhi NCR region.
As far as occupation is concerned, India stands out globally. Anthropic states that 45.2% of tasks map to software-related roles, the highest share among countries analysed (ahead of Vietnam at 42.1% and Egypt at 39.2%). Common activities include software development, coding assistance, and educational tasks tied to technical learning.
Modifying existing code is what Claude does the most in India
To derive these occupational insights, Anthropic employs a privacy-preserving classification method that maps anonymised conversation transcripts to the O*NET task taxonomy and SOC occupation groups. The top 10 most common O*NET tasks performed by Indian Claude users further illustrate this trend.
– Leading the list is “Modify existing software to correct errors, allow it to adapt to new hardware, or to improve its performance” (10.5% share).
– This is followed closely by “Write new programs or modify existing programs to meet customer requirements” (7.1%).
– Other high-ranking tasks include modifying software for hardware adaptation or interface upgrades (4.6%), designing/building/maintaining websites (3.0%), and assisting students with extra coursework help (1.8%).
– Tasks related to writing and maintaining programs, troubleshooting hardware/software issues, editing for readability, career instruction, and developing machine learning algorithms also feature prominently in the top 10.
These results clarify that Indian users primarily leverage Claude for core software engineering, debugging, web development, and technical tutoring — activities that directly support the country’s vast IT and services export economy.
Claude used to speed up work by 15 times
Anthropic states that Indian users demonstrate exceptional engagement with Claude. They achieve a 15x productivity speedup — completing tasks in an average of 14.8 minutes with AI versus 3.8 hours without — surpassing the global average of 12x. They also delegate higher levels of autonomy to the model (3.60 on a 1–5 scale, compared to 3.38 globally) and tackle more complex tasks less likely to be fully completable by humans alone (84.6% completable vs. 87.9% global).
Work-related use dominates at 51.3% (vs 46% globally), with coursework at 20.9% and personal tasks lower at 27.8% (vs 34.7% global). Indian prompts and AI responses reflect high education levels, ranking in the top 10% globally.