Genie Code is for building things. Apps, dashboards, and really any data products you want to communicate with.

Genie Agents are for insights on your data. What happened? Why did it happen? How did it happen? Will it happen in the future?

I want to share a story with you about a data product I built recently…I needed to share a point of view with a company on how Databricks was differentiated for Data + AI. The point — -> quality outputs are dependent on business context. In this exercise I used synthetic data from the company’s industry combined with my own insight on how they sell their products, how finance reports on revenue, and the unique language of their business (acronyms, product names, channel names, etc). As a business user, I was able to prompt Genie Code and construct an App and a Genie Agent to answer questions from a sales or finance professional.

Testing for quality and context is essential, and Genie Code was very helpful at flagging issues with my data. As a result, we added additional data tables which allowed us to explore a larger range of questions and provide deeper analysis.

My client said the demo really helped with understanding Databricks, Genie, and how AI solutions can be designed to provide more context, higher quality answers, and ultimately consistency (a trusted & certified data product). Genie Code made it really fast to develop the agentic prototype.