If you want an early look at how AI agents might reshape a workplace, a company that has rebuilt its own work around them is a reasonable place to start. Anthropic has published a report on its own workforce, drawing on internal data on hiring, AI use, output, calendars and hiring decisions, plus 17 employee interviews. The headline finding concerns the jobs people usually start in.
Junior roles are shrinking. Since early 2025, juniors fell from 21% to 14% of employees in Anthropic's technical departments, and their share of technical hires dropped faster, from 27% to 9%, with postings falling sharply since December 2025. By April, engineering and research teams had deprioritised lower-level technical hiring, and around a third of machine learning managers and a fifth of software engineering managers said they were hiring less because of AI. Administrative roles that usually grow with a company have barely grown, as staff build agents for scheduling, triage and purchase reviews.
The agents now do most of the hours. Agent work hours passed human hours in March 2026, and by August interactive agent hours outnumbered employees six to one. Output rose with them: code commits per engineer were five times their level of six months earlier. Following 192 engineers who adopted coding agents against 74 who had not, the amount of code added rose sharply after adoption, and sales representatives closed deals with more clients after adopting agents. The work changes shape too. Reading everything their agents produce would take even non-technical staff half their week, so people turn to building dashboards and monitoring pipelines, and the median employee spends 20% more time in meetings than a year ago. Hiring managers now stress autonomy and judgement over narrow technical skill.
For other employers, the fall in junior hiring is the part to watch. If entry-level work is the first to move to agents, the usual route by which people learn a trade narrows, and companies that rely on Claude Code or similar tools will face the same question about where their next senior staff come from. Anthropic says it has set up an early warning system for jobs at risk and has prioritised moving people internally; staff with a year's tenure were 50% more likely to move departments each month in 2026 than in 2025.
The caveats are real, and the authors note several. Anthropic grew very fast in this period, which shapes hiring on its own. More code does not mean better code; engineers who adopted fully autonomous agents were already on a steep upward path. And a company that sells AI models has unusual incentives and unusual access to them. Treat this as a vivid case study, not a forecast for typical workplaces.