AI agent

An agent receives a task – such as “Research three suppliers and create a comparison table” – and decides itself which steps are needed: searching, reading web pages, calculating, writing files. After each step it checks the result and plans ahead.

Agents can take over a lot of routine work but also make mistakes that add up over several steps. Clear limits, restricted permissions and human approval for important actions are therefore essential.