Language models & chatbots
An AI agent is a system that uses a language model to plan and carry out multi-step tasks on its own, deciding which actions and tools to use along the way.
A chatbot answers one message at a time. An agent is given a goal and works toward it. It breaks the goal into steps, takes an action, looks at the result, and decides what to do next. It might search the web, read a file, run code, or call another service, looping until the task is done or it gets stuck.
The language model is the agent's brain, and tool calling is its hands. Around them sits software that keeps track of progress, enforces limits, and asks a human when unsure. The word 'agentic' describes AI that acts this way, rather than only replying.
Agents are powerful and fragile. Each step can go wrong, and errors compound over long tasks. As of 2026, they are used for coding, research, and customer workflows, usually with a human able to review and approve important actions. Good design keeps them on a short leash.
You tell an agent 'book me the cheapest morning train to Manchester on Friday,' and it searches timetables, compares prices, and comes back with a ticket ready for you to confirm.