Glossary · AI & Agents
Function Calling / Tool Use
Overview
How tool calling works
Developers describe available tools to the model as structured schemas, each with a name, purpose, and parameters. When a request needs one, the model does not run the tool itself; it returns structured JSON naming the tool and its arguments. Your application executes that call, whether hitting an API, querying a database, or running code, then passes the result back so the model can finish its response.
Tool calling, agents, and MCP
Tool calling is the foundation of AI agents: an agent loops through calling tools, reading results, and deciding the next step until a task is done. The Model Context Protocol (MCP) standardizes how models connect to tools and data, so integrations can be reused across systems. Together they let AI act on live information instead of relying only on training data.
Why it matters for automation
Tool calling is what turns a chatbot into something that gets work done, booking appointments, updating a CRM, checking inventory, or sending messages. Our team uses it across AI Agents, Workflow Automation, and AI Voice Agents projects to connect models safely to real business systems, with permissions and guardrails around any action that changes data.
Where we use it
Related Zen in Tech services
How our team puts Function Calling / Tool Use to work in real projects.
FAQ
Function Calling / Tool Use — common questions
What's the difference between function calling and tool use?
They refer to the same capability. 'Function calling' is OpenAI's original term and 'tool use' is Anthropic's; both describe a model returning a structured request for an external function that your application then executes.
Is tool calling the same as an AI agent?
No. Tool calling is a single capability, a model requesting one function. An AI agent uses tool calling repeatedly in a loop, reasoning about results and choosing next steps to complete a multi-step goal.
Which AI models support tool calling?
Most modern LLMs support it, including OpenAI's GPT models, Anthropic's Claude, and Google Gemini. Support and reliability vary by model, so tool-heavy applications are usually tested and evaluated per provider.
Keep exploring
Related terms
Need Function Calling / Tool Use done right?
Book a free consultation and we’ll map the fastest, most cost-effective path for your project.
Knowledge hub
From our knowledge hub
All articles →15 Real AI Automation Examples You Can Deploy in 2026
15 practical AI automation examples across sales, support, marketing and operations, with the tools and ROI behind each. Get ideas you can ship now.
Read · 10 min →AI AutomationAI vs Traditional Automation: What's the Difference?
AI vs traditional automation: how rules-based workflows differ from AI-driven ones, and when each is the right tool for the job.
Read · 7 min →AI AutomationAI Agent vs Chatbot: What's the Difference?
AI agent vs chatbot: how they differ in autonomy, tools and tasks, when to use each, and how they work together. A clear side-by-side comparison.
Read · 7 min →