What Is an Agent?

Agents

Part of: AI Agents & Tool Use

Ask a plain chatbot "what's the weather in Tokyo right now?" and it will confidently invent a number. Ask an agent the same thing and it will pause, call a weather API, read the real answer, and tell you. Same model underneath. The difference is that the agent is allowed to do things , not just talk. What it is An agent is a language model wrapped in a loop that lets it take actions in the world and react to the results. A plain LLM is a one-shot function: text in, text out, done. An agent adds three pieces around that function, a set of tools it can call, a loop that runs until the job is finished, and memory of what it has done so far. The crucial shift: a chatbot only ever produces text. An agent produces text and decisions about which tool to run next. It can fetch a webpage, query a database, run code, or send an email, then feed the result back to itself and keep going. How it works Every agent cycles through the same four moves: 1. Observe. It reads the current state, your goal plus everything that has happened so far. 2. Decide. It predicts the next action: either call a tool, or produce a final answer. 3. Act. If it chose a tool, the surrounding program actually runs that

Challenge: The Tool Dispatcher