Function Calling

Full definition

Function calling was popularized by OpenAI in June 2023 and is now standard across Anthropic, Google, and open models. The model is fine-tuned to recognize when a user request maps to a tool, fill arguments from the conversation, and return a parseable structure. It is the substrate beneath tool use, agents, and structured-output applications. Pitfalls: hallucinated arguments, schema drift, and mishandling of parallel calls.

Why it matters

Function calling is the contract between probabilistic LLMs and deterministic enterprise systems. Done well, it lets LLMs participate safely in databases, payments, and workflows. Done poorly, it is a common cause of agent failures in production.

Example

A finance copilot exposes get_transaction(id) and refund_transaction(id, amount). The LLM resolves "refund the duplicate charge from yesterday" into two function calls; the application enforces the actual authorization.

Next steps

• Get a free workflow map →We map one of your workflows and give you a written verdict: software, a process change or AI. No obligation.
• AI consulting cost guide →Hourly rates and project budgets by firm tier, from independents to the Big Four.

Related terms

• Tool Use (LLM)
  LLM tool use is the pattern in which a language model acts through external functions, APIs, databases or code execution: the model requests a tool, the host application runs it and returns the result, and the model continues, turning a text generator into a controller that can act. Function calling is the structured-output mechanism most tool use is built on.
• Agentic Workflow
  An agentic workflow is a multi-step AI process where a model plans, takes actions through tools, observes results, and iterates toward a goal — replacing single-shot prompting with a loop that reasons, acts, and self-corrects.
• AI Agent / Agentic AI
  An AI agent is an LLM-driven system that, given a goal, plans a sequence of steps, calls tools or APIs, observes the results, and iterates until the goal is reached — going beyond single-turn chat to multi-step autonomous action.
• AI Guardrails
  AI guardrails are the layered controls — input filters, output classifiers, policy engines, schema validation, and structured generation — that constrain an AI system to safe, on-policy, and on-task behavior.

More deployment & operations terms

• Multi-Agent System
  A multi-agent system orchestrates multiple specialized AI agents — often LLM-powered, with distinct roles, tools, and prompts — that communicate, coordinate, and divide labor to solve a problem larger or more diverse than any single agent could handle reliably.
• Dense Passage Retrieval (DPR)
  Dense Passage Retrieval (DPR) is a retrieval method that embeds queries and passages into the same vector space using two BERT encoders (a dual encoder) trained with contrastive loss on (question, positive-passage) pairs.
• Embedding Model
  An embedding model is a neural network that maps text, images, or other inputs into fixed-dimensional vectors where semantic similarity corresponds to geometric closeness (cosine or dot-product distance).