Tool use: Tool use (also called function calling) is the ability of an AI model to request that an outside tool be run, such as a web search, a calculator, a database query, or a calendar lookup, and then use the result in its answer.
A language model on its own can only produce text. Tool use gives it hands. The application tells the model which tools exist and what inputs they take. When the model decides a tool would help, it outputs a structured request, for example "look up order 48213." The application runs the real lookup, hands the result back, and the model continues with actual data instead of a guess.
This is the mechanism behind most useful AI features you see: an assistant that checks the weather, runs code to compute a total, searches your files, creates a ticket, or sends a message. It is also the foundation of AI agents, which chain many tool calls together to complete multi-step tasks. Standards such as the Model Context Protocol make it easier to connect the same tools to different assistants.
Tool use shifts where errors happen. The model can pick the wrong tool, pass a wrong argument, or misinterpret the result, and a tool that takes actions can cause real-world consequences. That is why well-designed systems separate read-only tools from tools that change things, and require human approval before the second kind runs.
Example at work
A logistics coordinator asks an assistant, "Which of today's shipments to Texas are at risk of missing delivery?" The assistant calls a tracking tool for each open shipment, calls a weather tool for the routes, and then writes a short list with reasons. It has a "draft customer notice" tool but no "send" tool; the coordinator sends.
Why it matters
Tool use is what turns a chat window into something that can look up your data, do math correctly, and take actions on your behalf. Understanding it helps you judge what an AI product can really do, and helps you decide which actions should require your approval.
Related terms
- AI agentAn AI agent is a system that uses an AI model to pursue a goal on its own by planning steps, using tools such as web browsers, files, and software APIs, and adjusting based on results, rather than answering a single prompt.
- Model Context Protocol (MCP)The Model Context Protocol (MCP) is an open standard, introduced by Anthropic in late 2024, that lets AI assistants connect to external tools, data sources, and services through a common interface instead of custom one-off integrations.
- Agentic AIAgentic AI describes AI systems designed to act with autonomy: they take a goal, break it into steps, use tools, and carry the work through with limited human direction. It is the broader approach; an AI agent is a specific system built that way.
- API (application programming interface)An API (application programming interface) is a way for one piece of software to talk to another. An AI API lets your own tools, spreadsheets, or automations send a prompt to a model and receive the response, without a person typing in a chat window.
- Human-in-the-loopHuman-in-the-loop is a design approach in which a person reviews, approves, or corrects an AI system's output at defined points before it takes effect, combining the speed of automation with human judgment and accountability.