What changed
AI agents are moving into companies as named, role-based coworkers rather than staying in the background as software tools. A January poll of 1,261 managers found that 22% said their organisations had already added agents to corporate org charts, while Microsoft, Anything and others pushed products designed to work across email, Slack and other workplace systems.
Why it matters
The important shift is cultural as much as technical. These systems are being given names, roles, avatars and cloud computers so people can treat them like colleagues. Skydive agents, launched by Anything in August, can work across Slack, email and iMessage; Microsoft’s Scout handles jobs such as rescheduling meetings and drafting emails.
That packaging may make adoption easier, but it also blurs responsibility. At Pronto Housing, agents write contracts, schedule meetings, follow up with customers and help with coding. Employees have started saying they “worked with Alice” on a project. That sounds harmless until an agent makes a bad call and nobody is sure whether the mistake belongs to the system, its manager or the company that deployed it.
The commercial promise is obvious: an always-on digital worker that can absorb repetitive tasks. The harder question is whether these agents can reliably fit into human teams at scale. Their ability to imitate communication styles and improve through corrections may help them blend in, but it could also make their limitations less visible.
My view is that the workplace is adopting the metaphor of a coworker faster than it is settling the rules of employment around one. Companies may soon have org charts full of agents, while still improvising the basics: who reviews their work, who can overrule them and who answers when they fail.
Comments
No comments yet.