Databricks AI/BI Series: AI/BI Dashboards

Introduction

The Databricks Data + AI Summit this year had many new features and concepts revealed by the Databricks team. One that caught our eye in was the rebranding of “Lakeview Dashboards” to “AI/BI Dashboards”. Another one that also got our attention was the bigger reveal of “AI/BI Genie”, previously known as “Genie Spaces”. While both products are still fairly early in their lifecycles, they present different opportunities for organizations now and in the near future. Today, let’s take a quick look at what AI/BI Dashboards mean for your business. In a follow up article, we will cover AI/BI Genie.

A Brief Conversation About Dashboards

Dashboards are tools that help users visualize data, with these visualizations enabling decision making, at least, it is what they are intended for. There are plenty of time-proven dashboard vendors out there with tools such as Power BI and Tableau, as well as many promising new-comers such as Sigma, Omni, and Astrato. The question then turns to figuring out what place, if any, should Databricks’ AI/BI Dashboards have in your stack.

AI/BI Dashboards

What They Are Not :

The best place to start is to understand what AI/BI Dashboards are not intended to be:

  • In contrast to tools such as Power BI that can connect to a vast range of platforms and files such as Excel and CSVs, Databricks’ offering is not intended to bring data outside of the Unity Catalog’s realm.

  • Not intended to have every bell and whistle that dedicated BI tools like Sigma have. 

What Is Their Value ?

So, if AI/BI Dashboards are not intended to be a head-on competitor to these other well-known dashboarding tools, where does this Databricks offering fit, if anywhere in your organization? I am glad you asked:

  • Like with other dashboarding tools leveraging a live-connection, you pay for the cost of the compute used to retrieve and load the data. Unlike the vendors mentioned though, there are no licensing fees or additional subscriptions to pay to use Dashboards, which in many cases translates to cost savings to your organization.

  • What’s better than a great UI for creating charts like other tools have? Barely having to use the UI to create charts at all by leveraging AI to create charts using simple English. The text-to-chart capabilities are easy to use and smart, thanks to the connection to the Unity Catalog’s data governance capabilities. Ultimately, this means developing dashboards is easier and faster.

  • There is no need to bundle additional tools. AI/BI Dashboards are available right from the start.

Current Limitations:

While AI/BI Dashboards are now GA, as we mentioned early, they are still at an early stage. With that comes some of the follows pains that we expect will be worked out in the near future:

  • Tableau and Sigma have fantastic chart customization capabilities that make the dashboard look beautiful, but Databricks’ offering is still behind in this area. Chart formatting options need tender loving care as the offering matures.

  • The general canvas experience needs some adjustments to make it more dashboard developer friendly.

Thoughts & Conclusions

  1. As-is, we believe that AI/BI Dashboards satisfy basic reporting needs without having to bundle additional tools/vendors. Part of the value proposition of having a data intelligence platform.

  2. In the next 6 - 12 months, we expect reporting capabilities will be suitable for mid-management level reporting and after that, we expect to start seeing executive level quality reporting become possible.

  3. Tools like Power BI, Tableau, and Sigma will still have their place inside of the Databricks ecosystem and your stack, but they will become less prevalent for companies with less complex reporting needs and/or businesses looking to run their stack as lean as possible.

We hope you found this article helpful. Below you will find a video that we did creating a simple dashboard heavily leveraging the text-to-chart capabilities. Stay tuned for our article covering AI/BI Genie, which aims to be the middle ground between your data and LLMs to help you answer data questions that you would normally not be able to answer using dashboards.

Watch a Quick Walkthrough

For a more in-depth look at these features, check out this 2’ video:

Previous
Previous

Databricks Myths vs. My Own Personal Experience

Next
Next

Evaluating Databricks' Cost Control Features: A Closer Look at Budgets and Cost Dashboard