Anthropic's staff is skewing more senior as meeting time climbs

Anthropic published a report on Friday describing how its own workforce has changed over the past year, and it says the shift could hint at how artificial intelligence will reshape hiring and work organization more broadly. The company says it has “thoroughly molded its workflows around AI,” and argues its internal data offers a preview of trends that could spread to other employers, even if few adopt AI as thoroughly as Anthropic has.

The clearest change is in who gets hired. The proportion of junior-level employees in Anthropic’s technical departments fell from 21% to 14% since early 2025, the report says. The drop is sharper among new hires: the share of junior technical recruits fell from 27% to 9% over the same period. Anthropic’s report does not spell out a single cause for the decline, but the numbers describe a company increasingly staffed by more experienced technical employees.

Technical roles also make up a smaller slice of Anthropic’s overall workforce, down from 61% to 56%. The company attributes this to the “natural expansion of business functions as the company scales,” and to coding tools that let non-technical employees build their own software, cutting the need to route every project through an engineer.

The report also describes a workplace where managing Anthropic’s own Claude models has become a job in itself. In an April 2026 survey cited in the report, 27% of employees said Claude had made them more productive, while 20% said their role had shifted “from actually doing work to managing Claudes.” One engineer compared the experience to childcare, saying “every child brings to you concerns with an equal demand for your attention and the various [Claudes] also do that.” The report frames this as evidence that human attention and judgment, not computing power, are becoming the limiting factor on productivity, as Claude produces a volume of text that still needs a person to read and organize it.

Anthropic credits AI adoption with gains including more code written per engineer, more sales closed per rep, and more contracts processed. The company adds a caveat of its own: “more code does not necessarily mean better code or business impact,” though it says it has not found evidence that code quality is slipping.

Those gains have not cut down on meetings. Calendar data cited in the report shows meeting time per employee is up 20% compared with a year earlier, and up 16% among technical staff specifically. Anthropic attributes the increase to employees’ reliance on “human trust and accountability,” with workers describing face-to-face coordination as something they still can’t do without. The report does not say whether fewer junior hires, more oversight of Claude, and more meetings will continue as Anthropic’s AI tools keep changing, or whether other companies adopting those tools will see the same pattern show up in their own hiring and calendars.