Introducing Databricks Assistant for Help

Building on the momentum of Databricks Assistant, the context-aware AI assistant integrated within Databricks Notebooks, SQL editor, and file editor, and now powering AI-generated documentation in Unity Catalog, we are thrilled to announce the public preview release of Databricks Assistant for Help.

Databricks Assistant for Help is the new way to learn, explore, find, troubleshoot, and get support using AI. Users will be able to leverage AI to search across Databricks documentation to answer questions you may have about functionality and errors. It can also help you get in contact with our support team if additional assistance is needed and you have support access. The Databricks Assistant for Help is accessible from any page and can be found in the top navigation bar next to the username.

Just ask questions 

You can ask the Assistant questions to learn about the product and get help with Databricks-tuned knowledge. The structure of these questions can be simple, descriptive, or conversational.

Here are some examples of the types of questions you can ask.

Simple questions:

“How can I view cluster metrics in realtime?”

Descriptive questions

“We scheduled a Databricks job to run a complex ETL process, but it failed with an error message indicating a resource constraint. The job involves heavy data transformations and joins, resulting in high memory consumption. We would like to address the issue and ensure the successful execution of the job.

What could be causing the resource constraint error and how can we resolve it?”

You could follow up on this question with “Are there any Databricks configuration settings that can be adjusted to optimize memory usage for such jobs?”

Conversational questions

You can ask conversational questions. For example: “What is Unity Catalog?”, and then “How do I enable it?”

You can even get support

If you can’t find what you are looking for in the Databricks Assistant for Help, try asking it for support. Customers with support contracts will be linked to an in-product ticket submission experience. Submitting tickets through this experience is faster and includes richer data than the external ticket submission experience. Users without support contracts will be linked to helpful external resources. 

Setup

The Databricks Assistant for Help can be enabled together with and uses the same services as the Databricks Assistant. Refer to the Databricks Assistant FAQ to learn how to enable it in your workspace. Try out the Databricks Assistant today and let us know what you think via the in-product feedback link.

Finally, don’t forget to help us help you get better answers. You can do this by clicking the thumbs up / thumbs down for the answers you receive. 

This way Databricks Assistant will get better the more you use it.