THE CRUNCH
At the AI startup Anything, founder Dhruv Amin works alongside Gwen, an executive assistant, Cody, a coding partner, and Justin, an office manager who stocks the kitchen. None of them are people. They are AI agents, software that handles tasks without human intervention, and Anything now has roughly 150 of them alongside 15 human employees. Amin believes that ratio will soon be the norm, telling the Guardian that within a year every company will have more AI co-workers than human ones.
The trend is visible beyond one startup. Viktor sells an "AI chief of staff", Lindy makes customisable agents for sales, support and admin work, and OpenAI, Anthropic and Microsoft all offer agent products that plug into thousands of other tools. At Lindy, which laid off a third of its staff and now has about 20 employees, senior growth marketing manager Haneen Azhar says agents handle product announcements, social posts and marketing videos that used to be done by around 100 people in the company's creator programme.
The adoption is spreading past the San Francisco bubble. McKinsey's State of AI Global Survey found about 40% of large organisations used AI agents in some capacity between May 2025 and June 2026, up from 27% the year before, and Boston Consulting Group reports companies expect to shrink workforces by 10% to 15% as they delegate more to AI. Yet MIT research indicates companies are not currently seeing major productivity gains from AI.
The human cost is already showing. Agents are alienating customers and colleagues who want human interaction, eliminating roles, and turning employees into overnight managers of bots that often make mistakes. Azhar herself worries about the effect on newcomers: "If I was trying to get my foot in the door right now, I'd be worse off than I was a year ago."