Listen: Tool Use
How a model acts beyond text: it emits a structured call to an external tool such as search, code, or an API; a harness runs it and feeds the result back. The mechanism under AI agents.
Transcript
Tool use, which you might also hear called function calling, is how a language model does more than just write text. It is the mechanism that lets a model interact with the outside world, using external tools like search engines, calculators, or database queries.
The model itself does not actually run these programs. Instead, it is given a list of available tools. When a task requires one, the model outputs a structured request. An external system executes that request and feeds the result back into the prompt.
This loop of calling, executing, and reading is what turns a simple text predictor into an active AI agent. It is the foundation of the entire agent stack. New standards, like the Model Context Protocol, are making it easier to connect models to these tools and websites.
This shift is redefining the web. In the future, brands will not be reached through a page a person reads, but through a tool an AI agent calls behind the scenes.