Function Calling (Tool Use)
Function Calling is the capability of LLMs to call external functions or APIs as part of generating responses. It enables agents to take actions, query databases, or access real-time information beyond the model's training.
Definition
When an LLM needs to perform a calculation, look up current information, or trigger an action, it can call a function defined by the system. The function executes, returns a result, and the LLM incorporates that result into its response. This breaks the model's dependency on training data and inference only, enabling real-time, dynamic responses.
Function calling is essential for agents and agentic workflows. An agent researching prices calls a price lookup function, an agent writing an email calls a send email function, and an agent searching the web calls a search function. The model decides which function to call and with what parameters based on context.
Why it matters for AI visibility
Function calling means AI systems can access your APIs, check current information, and interact with your product directly during answer generation or agent workflows. Your brand's visibility benefits from well-documented APIs that agents can discover and use reliably to access current product information.
Related terms
AI Agents
AI Agents are autonomous systems that use LLMs to perceive their environment, make decisions, and take actions to accomplish goals. They can use tools, run code, and execute multi-step workflows without human intervention.
AIAgentic Workflows
Agentic Workflows are automated processes orchestrated by AI agents that execute multi-step tasks, make decisions between steps, and adapt to outcomes. They combine language understanding with action execution.
AIModel Context Protocol (MCP)
MCP is an open protocol introduced by Anthropic for connecting AI models to tools, data sources, and external systems in a standardized way. It enables safe, controlled access for agents and tools.
AIComputer Use
Computer Use is the capability of AI systems to see and interact with computers through screenshots and simulated input, enabling them to navigate websites and use software tools directly.