Native function calling
Agent Studio runs all agents and workflows on native function calling, also known as tool calling. This replaces the traditional text-based parsing methods to achieve deterministic, reliable execution of enterprise AI agents.
Agent Studio provides the model with a structured definition of each available tool, including its name and parameters. Instead of returning raw text, the LLM responds natively with a structured tool-call object.
This shift to native function calling eliminates the need for complex text parsing, allowing arguments to arrive directly as structured data.
Key benefits and capabilities of native function calling
- Schema enforcement — Enforces parameter alignment against a strictly defined schema to guarantee high accuracy and ensures that the outputs match your expected data types.
- Parallel processing — Enables the simultaneous invocation of multiple tool calls within a single turn to significantly increase operational efficiency.
- Resource optimization — Eliminates malformed-action retries through strict schema adherence to directly reduce both latency and operational costs.
