‘An army of Claudes, with every one asking for my attention’: Anthropic calls human attention the bottleneck in the botsitting world to come

AI agents are now moving so fast that 80% of the elapsed time in Claude sessions was due to the agents waiting on a human’s input, Anthropic’s report found. “This acceleration makes human attention the bottleneck for agent output,” the report read.

Instead of actively doing the work themselves, even non-supervisor employees are increasingly acting like managers to multiple AI agents that complete tasks but also require direct attention at times.

One employee described the experience of bot management as “an army of Claudes, with every one asking for my attention.”

That bottleneck is the flip side of a big jump in output. Employees are producing more and moving faster, the company found. AI agents collectively surpassed human working hours for the first time in March and by August, AI agent hours worked outnumbered the hours worked by humans six to one.

Software engineers’ code commits have also increased five times over the past six months. And even non-technical employees have personal agents monitoring their email and calendars, sending to-do lists compiled from messages, and briefing them on the purpose of their upcoming meetings.

These changes have taken some manual labor off employees’ plates, but they’ve also transformed what the workday looks like.

Managing agents is becoming a job in itself. The share of employees’ interactive Claude sessions devoted to this kind of “meta-work” rose from 10% to 22%, according to the report. Even for non-technical employees, reading everything their Claude agents produce would take up about half their work week. And private work, done in unlogged meetings and direct messages, is “invisible, confusing and discouraged in an AI-heavy workplace,” the report said.

The median employee now spends 20% more time in meetings than a year ago. Part of that is the company’s growth (Anthropic now has more than 3,500 employees), and part is an emphasis on hiring in non-technical roles. But meeting time is up 16% even among technical individual contributors, a sign that humans are spending more of their time on coordination and decisions while agents handle tasks like repetitive coding on their own.

And the pace keeps ratcheting up. As soon as a new model is available, Anthropic employees start tackling harder problems, the report found. One employee described the “eternal pressure at Anthropic to move as fast as possible and use Claude for everything.”

Jobs redefined

In some cases, employees using AI have streamlined their own responsibilities away. The report mentions a product design team, which previously wrote short pieces of copy for the company’s website, including greetings displayed on Claude.ai, started using AI to write the copy this Spring.

“The team then rotated into agent prompting and began writing model evals and prompts for Claude,” the report read.

This phenomenon is increasingly common at Anthropic as AI takes over tasks. The company claimed in the report that employees working at the company for at least a year were 50% more likely to move across departments each month in 2026 than in 2025. Still, the company said in the report it has made an effort to identify at risk jobs and reassign employees to other positions. 

The hiring process is changing, too. Hiring managers increasingly emphasize applicants’ autonomy and judgment relative to specific technical expertise, the report said.

And there are fewer entry-level people being hired. The share of junior-level employees in technical departments fell to 14% from 21% in early 2025, and the share of junior technical hires fell even faster, from 27% to 9%. Open positions for those junior roles also plummeted since December, the report found.

That’s likely to add to anxiety about AI job displacement, which is running especially high among young people. A survey from recruiter platform iCIMS published earlier this year found that 51% of Gen Z respondents considered AI the greatest threat to their job security.

A Deutsche Bank Research survey from last year also found that nearly one in five Gen Z workers is worried AI will cost them their job over the next two years, compared to just one in 10 Baby Boomers who said the same. 

The bigger picture

In the future, employees at Anthropic see AI taking up more and more of their daily responsibilities. As of August, employees at Anthropic said they expect AI to be able to complete three-quarters of their work by 2030, compared to a less optimistic 28% predicted by other U.S. workers, according to another study by Anthropic.

It’s unclear whether the changes at Anthropic are an accurate reflection of how the average workplace might look in the future. To be clear, the AI company has massive resources and is nearing a potential $2 trillion IPO that could break records. The company is also implementing AI deeply in its workflows to an extent other companies may not be.

Still, the report paints a picture of transformative change that could reflect how companies and corporate roles change as AI adoption takes root. 

As tasks like writing code, identifying bugs, or writing emails get automated away by AI, employees at Anthropic have now found that the big picture tasks are increasingly what count. 

As one employee in the report put it, “that 20% that was strategic becomes 90% of the job.”