What Is Botsitting?

As AI agents become more capable, a new term is emerging for the humans supervising them: botsitting.

The name is a play on babysitting. Rather than completing a task yourself, you give it to an AI agent and monitor what happens, stepping in if it gets stuck, makes a mistake or encounters something that requires human judgement.

Unlike a chatbot, which generally responds to individual questions or prompts, an AI agent can carry out multiple steps towards a goal, potentially using tools, interacting with other systems and making decisions along the way. That gives it more freedom to act on our behalf, but it also creates a need for someone to make sure everything is going as expected.

Botsitting describes that role. The AI might carry out most of the task, while a person checks its progress, approves important actions or takes over when something needs human judgement.

Why Does It Matter?

As AI agents take on more work, our role may increasingly shift from completing every step ourselves to overseeing what AI is doing for us. That still requires people to understand the task well enough to recognise when something has gone wrong and know when they need to intervene.

Botsitting might be a slightly silly name, but the idea behind it is simple: as we give AI more freedom to act, someone still needs to keep an eye on it.