The composition of Anthropic’s workforce and the nature of work across the company have changed significantly over the past year, according to a report published by the artificial intelligence firm on Friday.
The company said its workplace trends could offer insights into how artificial intelligence may reshape work organisation, recruitment and labour demand across the wider economy.
Anthropic noted that its operations have been extensively structured around AI, making its workforce a potential indicator of how businesses could adapt as the technology becomes more integrated into everyday work.
Anthropic has cut its junior-level technical workforce, with the proportion of junior employees in technical departments falling from 21 per cent in early 2025 to 14 per cent.
The share of junior recruits in technical roles dropped more sharply, from 27 per cent to nine per cent over the same period.
The company also reported that technical teams now account for 56 per cent of its workforce, down from 61 per cent.
It attributed the decline to the expansion of other business functions as the company grows and the increasing use of AI-powered coding tools, which enable non-technical employees to develop software.
In an April 2026 survey, 27 per cent of Anthropic employees said Claude had improved their productivity, while 20 per cent reported that their roles had shifted from performing tasks themselves to managing multiple AI agents.
One engineer compared the experience to babysitting, explaining that each AI agent demands attention in much the same way as children seeking help.
The report also warned that human attention and judgement are increasingly becoming bottlenecks to productivity, as Claude generates large volumes of text that employees must read, review and organise.
Anthropic said the adoption of artificial intelligence has improved productivity across its operations, citing increased code production per employee, more sales per representative and faster contract processing.
However, the company cautioned that producing more code does not automatically translate into better software quality or stronger business results.
Despite this, Anthropic said it had found no evidence that code quality was deteriorating as employees increasingly relied on AI tools.
