In the era of agentic AI, data quality and accessibility drive agentic transformation. The partnership between Informatica and Databricks empowers organizations to build and deploy enterprise-grade AI applications on a foundation of trusted, high-quality data. A key component of this collaboration is the Informatica Databricks Agent Bricks Connector, a powerful tool streamlining the development and deployment of enterprise-grade AI applications built with Agent Bricks. Let's delve into its technical capabilities and explore how our pre-built recipes can accelerate your journey to become an Agentic Enterprise.

The Informatica Connector for Databricks Agent Bricks bridges Informatica's Intelligent Data Management Cloud (IDMC) and the Databricks Agent Bricks services. It’s more than just a simple connector; it’s an integration tool that enables you to interact directly with Databricks' advanced AI capabilities from within the Informatica platform. The Connector natively supports OAuth 2.0, providing a secure and standard way to handle authentication, strengthening data security by eliminating the need to manage and share sensitive credentials. You can leverage advanced LLM inference, perform RAG-based searches and query your data in natural language via Databricks Genie, with complete assurance of robust model governance.

To help you get started quickly, we've developed a series of pre-built "recipes" showcasing the power of the Connector for Databricks Agent Bricks n real-world scenarios. These recipes provide solutions that you can easily adapt to your own needs. Here’s how you can get started with recipes of your own:

1. Copy Recipe 
   Go to Application Integration > Recipes, and search for  Agent Bricks to see all Databricks Agent Bricks recipes. Click on a recipe, select Use and choose a folder to copy recipe assets. 



2. Configure the connections
   Configure and publish the connections (Databricks Agent Bricks Email, FHIR, etc.) available in the recipe.
3. Publish the processes and subprocess
   You can publish all the main and sub processes to use the selected recipe in your organization. Please follow the documentation for detailed steps for each recipe as shown in the image below:

The following sections will walk you through a few of these recipes.

Databricks Agent Bricks Chat with File

This recipe helps you to build a powerful "chat with your data" application. You can upload a file in various formats, then use a large language model (LLM) hosted on Agent Bricks to ask questions and get answers based on the content of that file.

Simple RAG Consumption with Databricks Agent Bricks

For those who want to build their own RAG applications from the ground up, this recipe provides a solid starting point. It includes the essential components for RAG workflows, using Databricks Vector Search and LLM inference, that you can customize and extend to meet your specific requirements.

Patient Food Vendor Recommendation using Databricks Agent Bricks

This recipe demonstrates a more complex, multi-step AI workflow. It extracts a patient's food allergy data from a FHIR server, searches a Databricks vector index to identify suitable food vendors, and then automatically emails a curated list of recommendations to the user. This example powerfully demonstrates how AI agents can automate business processes across various industries, including healthcare in this case.

The combination of IDMC, the Informatica Connector for Databricks Agent Bricks and our growing library of recipes, provides a powerful solution for any enterprise looking to activate its AI initiatives with Databricks Agent Bricks and gain real data intelligence. By simplifying the technical complexities of AI agent development and providing pre-built recipes for common use cases, we are empowering you to focus on what matters most: unlocking the value of your data in the Databricks Data Intelligence platform to drive your transformation into an Agentic Enterprise.