Google adds agentic workflow tools to Gemini API with Google Maps grounding

Google updated the Gemini API to allow combining built-in tools and custom function calling in a single API call, enabling multi-step agentic workflows without manual orchestration. New capabilities include cross-tool context circulation and tool response IDs. Google also expanded built-in tool support to include Grounding with Google Maps for the Gemini 3 model family, giving models access to spatial data, local business information, and location-aware responses.

Chaining built-in tools with custom functions in one call removes the orchestration overhead that makes multi-step agent workflows fragile. The Google Maps grounding is a concrete differentiation for applications requiring location awareness: travel, logistics, local search, and physical-world agent tasks. For developers already on the Gemini API, these updates reduce the glue code needed to build reliable agentic pipelines.