What it means
Tool calling is a capability that lets an AI model request the use of external functions, such as searching a database, checking a calendar or performing a calculation, by producing a structured request. The surrounding software executes the request and returns the result to the model; the model itself does not run anything.
Real-world example
A visitor asks a museum's assistant, "Is the exhibition on the Bauhaus open on Monday?" The model recognises it needs current information and requests the "get_opening_hours" tool with the date. The museum's system queries its calendar and returns the result, which the model then uses in its answer.
Why it matters for AI adoption
Tool calling connects AI to real systems and actions, which makes it powerful and also where many risks arise. Organisations decide which tools are available, what each tool may do, which inputs are checked and which actions need human confirmation.
Questions to ask
- Which tools can the AI call, and what can each one read or change?
- Are tool inputs validated, and do any actions require confirmation by a person?