Function Calling
An LLM capability to output structured requests to external functions or APIs, enabling AI to take actions like searching databases or executing code.
In-Depth Explanation
Function calling (also called tool use) is a capability that allows LLMs to output structured requests to external functions, tools, or APIs. This bridges the gap between AI understanding and AI action.
How function calling works:
- Define available functions with parameters
- User sends a request
- LLM decides if a function is needed
- LLM outputs a structured function call
- Your code executes the function
- Results are returned to the LLM
- LLM generates final response using results
Function definition includes:
- Name: What the function is called
- Description: What it does (helps LLM decide when to use it)
- Parameters: Input schema with types and descriptions
- Required vs optional parameters
Common function calling use cases:
- Database queries
- API integrations
- Calculations
- Web searches
- File operations
- System actions
This capability transforms chatbots into agents that can take real actions, not just provide information.
Business Context
Function calling transforms chatbots into capable agents that can check inventory in Shopify, process orders in Stripe, or query your Salesforce CRM in real-time - essential for US e-commerce and B2B operations.
How Clever Ops Uses This
We implement function calling for US businesses to create AI assistants that integrate with popular American platforms - checking orders in Shopify, updating Salesforce records, sending Slack notifications, and connecting to business systems.
Example Use Case
"AI determines it needs to check order status and outputs: {"function": "get_order", "order_id": "12345"} which your code executes against your database."