Anthropic Discloses Internal Shifts: Fewer Junior Technical Roles, More Time in Meetings

What the report says

Anthropic published an economic research report titled "How employment is changing at Anthropic," dated October 16, 2026 on the page. Anthropic says the report draws on internal data covering employment, AI usage, output metrics, employee surveys, internal communication patterns, employee calendars and hiring decision records. It adds 17 employee interviews to that data.

Anthropic says the share of junior staff in technical departments has fallen from 21% to 14% since early 2025. Over the same period, junior technical roles' share of new hires fell from 27% to 9%. Technical departments' share of total headcount fell from 61% to 56% over the past year. Anthropic attributes this partly to business functions expanding naturally as the company grows, and partly to coding tools possibly letting non-technical employees build software themselves.

AI agent hours and productivity

AI agents are AI programs that carry out tasks on their own once instructed, such as automatically fixing code bugs. Anthropic says that by August, the interactive AI agents that employees directed worked six times as many hours as the employees did. Human hours were estimated as headcount times 40 hours per week. Each engineer's code commits (the number of times code changes are saved into a project) were five times what they were six months earlier. Anthropic cautions that employees did not adopt the tools at random. If employees took them up just before large projects, estimates of productivity gains after adoption could be biased.

Anthropic acknowledges that "more code does not necessarily mean better code or business impact." It compared code written by agents with code written directly by the same employee, in the same month and in the same directory. It found no difference in how often the code was later deleted or reverted. It also notes that code quality has many dimensions and that AI errors may be harder to spot than human ones.

Human attention becomes the bottleneck

An interactive Claude session is one where an employee and Claude collaborate back and forth. Anthropic says about 80% of the time in these sessions is spent idle, waiting for human input. Even for non-technical groups, these sessions produce so much output that reading it all would take more than half a work week.

Anthropic calls one kind of work "meta-work". Instead of doing the task directly, employees build skills for AI agents, create plans and reviews for other agents, monitor processes and so on. Meta-work rose from 10% of the average employee's interactive Claude sessions in autumn 2025 to 22% in spring 2026. In an April 2026 employee survey, 27% of respondents described productivity gains from Claude tools. Another 20% described their role shifting from doing the work to managing Claude.

Based on calendar data, Anthropic says the median employee spends 20% more time in meetings than a year ago. For technical individual contributors (employees who do hands-on technical work without managing a team), the increase is 16%. Anthropic says part of the increase reflects company growth and a rise in meeting-heavy non-technical roles. Employees also repeatedly described face-to-face coordination as indispensable.

Hiring and the company's response

Anthropic looked at the feedback interviewers wrote about candidates. In 2024, about 5% of feedback for technical roles and 7% for non-technical roles mentioned a candidate's track record with AI tools. By 2026 that rate had tripled for technical roles and quadrupled for non-technical roles. In technical interview feedback, the biggest increases were in comments on autonomy and judgment, while domain expertise took up a smaller share of the assessment.

Anthropic says employees with at least a year of tenure were 50% more likely in 2026 than in 2025 to move across departments in any given month. The company is building an early-warning system to track roles facing AI-driven change, and it gives priority to internal transfers. Anthropic says these measures do not directly address the slowdown in junior hiring. It says action by employers must come alongside government policy on retraining and income support.

What this means for ordinary readers

The report describes how work is changing at a company that uses AI heavily. In Anthropic's account, the emphasis is moving away from doing the work yourself and toward managing agents, reviewing their output and coordinating with other people. For job seekers and current workers, the report says this company now gives more weight to autonomy, judgment and skill with AI tools when hiring. It remains the self-description of a single company.

Browse the latest news in this topic

• AI news