Real-World Data Shows Claude Boosting Productivity, Anthropic Says
Real usage data shows AI compressing 90-minute workflows into minutes, though experts warn the benefits vary sharply across industries.
Topics
News
- DeepSeek Pushes Open-Source Math AI to Olympiad Heights
- Telecom Firms Flag Gaps in DPDP Rules, Seek Sector-Specific Guidance
- SAP Taps TCS to Overhaul Internal IT and Cloud Stack
- Deepwatch Sets Up Bengaluru GCC to Power AI-Driven Cybersecurity Push
- Reliance, L&T Make $13.5 Billion Push Into AI Data Centers
- India’s Hardware Startup Boom Faces Roadblocks Despite $296 Billion Potential
Image source: Chetan Jha/MITSMR India]
Work requiring about 90 minutes of human effort is now finished roughly 80% faster when supported by Claude, Anthropic said in a new analysis of 100,000 real user conversations.
The study is part of the company’s new Economic Index, which tracks how AI is reshaping everyday work by analysing real usage rather than hypothetical scenarios.
Anthropic used a privacy-preserving method to estimate how long tasks would take with and without AI help, offering one of the clearest measures yet of how much time large models may be saving across knowledge-heavy roles.
Anthropic said that if similar gains were applied across the US economy, annual labour-productivity growth could rise by as much as 1.8% over the next decade, or nearly twice the pace of recent years.
The company stressed the figure is not a forecast but an upper-bound estimate, noting that actual gains depend on adoption rates, verification time and organizational readiness.
The research shows that AI’s impact is highly uneven across professions. Claude is used most intensively for legal, managerial, education, research and media work, where drafting, reasoning and multi-step planning dominate.
These tasks show the steepest compression, often shrinking from over an hour to a matter of minutes. Healthcare administration also sees sharp gains, with documentation and assistance tasks accelerating by up to 90%.
Hardware, engineering and technical-support tasks see more modest improvements of about 50-60%, reflecting the mix of digital and physical work involved. Jobs that rely on real-world judgment or physical presence such as inspections or interpreting diagnostic images show little meaningful change.
Anthropic said its estimates do not include the time workers spend checking or refining AI output, which remains essential for accuracy and compliance in many industries.
The modelling also compresses long and short tasks toward the middle, meaning the extremes may be less precisely measured.
Even with these limits, the company said the findings offer a more grounded view of how AI is influencing day-to-day productivity.
The company plans to update the Anthropic Economic Index as usage patterns evolve and companies embed AI more deeply into workflows, aiming to map not just what tasks AI performs but how those tasks translate into measurable economic gains.
