How employment is changing at Anthropic

At Anthropic, rapid deployment of internal AI agents has fundamentally reorganized work: agent runtime now exceeds human work hours by six to one, and hiring for junior technical roles has fallen sharply (technical share of juniors dropped from 21% to 14%; junior hires from 27% to 9%). Agents handle coding, testing, scheduling and many clerical tasks, prompting managers to prioritize autonomy and judgment over narrow domain expertise. Non-technical employees increasingly use coding agents to build dashboards and pipelines, while traditional administrative roles have been automated or shifted to higher-order chief-of-staff work. As agents diffuse, employees are pressured to streamline themselves out of bottleneck tasks and rotate into roles focused on prompting, evaluation, and oversight.

Measured output rose across functions: code commits per engineer jumped fivefold in six months, lawyers processed more contracts, accounts payable moved invoices faster, and sales closed more deals. A within-person comparison of engineers who adopted coding agents shows sharp post-adoption increases in code output without higher rates of reverts or removed lines as a basic quality check. Coding-agent sessions for non-technical staff climbed from 2 to 90 per month in a sample, and employee redeployments rose (workers were 50% more likely to change departments in 2026 versus 2025). Anthropic has instituted an early-warning system to track roles at risk and prioritized internal redeployments, while workers increasingly perform “meta-work” to monitor and visualize agent pipelines because human attention has become the limiting resource.