Tool Use (AI agents)
Definition
Tool use is when a language model calls external functions—search, code execution, APIs, or browsers—to gather data or take actions while answering.
By Vinespire Editorial Team, Editorial ·
This term is part of the full AI search glossary.
Full definition
Modern assistants are not limited to next-token prediction. With tool use (sometimes called function calling or agents), the model selects a tool, receives structured results, and continues generating with that evidence. Web search, calculators, and CRM lookups are common examples.
Tool use changes AI search dynamics: a response may reflect live retrieval rather than training memory alone. Brands then compete to be found by the tools the agent invokes—indexes, maps, shopping graphs—not only by static model weights.
From a product perspective, tool use improves grounding and reduces some hallucinations; from a publisher perspective, it raises the value of crawl access, APIs, and well-structured public data. See Grounding, RAG, and Generative UI.
Example
An assistant answers “flights tomorrow to Austin” by calling a travel API, then formats prices and times in the chat response.