As we all know, Agentic AI is at the forefront of business transformation. Connecting to heterogeneous data sources is key to providing context from your Enterprise to the Gen AI models like Gemini, in order to reason across your data and provide incredible value.
I have created a demo showing how you can use ADK with Application Integration Connectors (to Cloud SQL (Postgres and SQL Server)) in addition to connecting with Vertex AI RAG Engine (Google’s managed RAG solution). Since Integration Connectors are tightly coupled with ADK’s Google Cloud Tools, it is incredibly simple to integrate them into your ADK Agents.
In this demo, you will use natural language to talk to your data from two databases and RAG Engine.
- NYC Taxi Ride Data (Cloud SQL - Postgres; a subset of data from a public dataset on Kaggle)
- Daily NYC Weather Data (Cloud SQL - SQL Server; a synthetic dataset providing weather conditions for the days represented in the NYC Taxi dataset)
- Car Models for Weather Conditions (RAG Engine; a synthetic dataset that provides recommended cars for weather conditions stored on various structured and unstructured file types)
It is amazing to see how Gemini 2.5 Flash (via the ADK Agent) can query disparate data sources that normally cannot be integrated without a lot of custom development. In addition, it reasons over this data and performs real-time joining of this source data - something not easily possible without the ADK Agent.
Feel free to take a look at the GitHub repository. All source data is included so you can try this for yourself! (This is a personal project and not endorsed by Google. It is just a demo, so use it accordingly. 
)