Quick answer: Wire LLMs to external tools + APIs — enables agentic behavior.
Tool/Function Calling is a technique that connects Large Language Models to external tools, APIs, and databases. Instead of an LLM only generating text, it can now decide when and how to call specific functions—like fetching data from an API, performing calculations, or querying a database—then incorporate those results into its response. For example, an AI assistant can call a weather API to provide current conditions, a calculator function to solve math problems, or a database query to retrieve customer information. This transforms LLMs from passive text generators into active agents that can take actions, solve real-world problems, and interact with live systems. Tool calling is the foundation of autonomous AI agents, chatbots that can actually complete tasks, and AI systems that can reliably access accurate, up-to-date information.