AI agent runtime at Anthropic has reached six times employee working hours. As AI takes on hands-on tasks such as coding and contract reviews, employees are shifting from doing work directly to managing and supervising multiple AIs.

Japan’s GIGAZINE reported on Oct. 19 that Anthropic on Oct. 16 released research results based on an analysis of internal hiring data, AI usage records, and employee surveys and interviews.

The research found that in March, total AI agent runtime exceeded human working hours for the first time. In August, runtime for interactive AI agents used by employees rose to as much as six times human working hours. In research, it reached 19 times.

Work volume also increased with AI adoption. Code changes per engineer rose 5.1 times from December last year. Query volume for data analysts and contract processing volume for legal staff rose 3.1 times each, and contracts handled by sales staff rose 1.7 times.

Anthropic said the company’s rapid growth may also have had an impact, making it difficult to conclude that all performance gains were due to AI adoption.

As AI workloads grew, new problems also emerged. About 80 percent of Claude session time spent conversing with employees was used waiting for human input. Non-technical staff were estimated to need more than half their weekly working hours to read all AI-generated documents.

Meta-work, in which employees instruct AI and check results, also increased. The share of related conversations rose to 22 percent in spring 2026 from 10 percent in autumn 2025. In an employee survey, 20 percent of respondents said their role had shifted to managing AI instead of working directly.

Hiring patterns also changed. The share of entry-level roles in technical functions fell to 14 percent from 21 percent, and the share of entry-level positions among new technical hires dropped to 9 percent from 27 percent.

Median meeting time among employees rose 20 percent from a year earlier. The analysis found that even if AI replaces hands-on work, humans are still needed for key decisions and coordination between departments.