Hi @rawalsparsh99,
Knowledge Assistant is part of the Agent Bricks feature set, and there are a few specific requirements that all need to be in place simultaneously for the feature tiles to become active. Since you mentioned you have admin rights, previews enabled, serverless budget quota configured, and a Premium tier plan, let me walk through the full checklist so we can narrow down what might still be blocking you.
REQUIREMENTS CHECKLIST
1. Mosaic AI Agent Bricks Preview must be enabled. This is a specific preview toggle (separate from other AI-related previews). Go to your workspace top bar, click your username, select "Previews", and confirm that the "Mosaic AI Agent Bricks" toggle is turned on. There is also an account-level preview setting, so your account admin should verify it is enabled in the account console (accounts.cloud.databricks.com) under the "Previews" section as well.
2. Serverless compute must be enabled for the workspace. This is separate from just having a serverless budget policy. Confirm under Admin Settings that serverless compute is turned on for the workspace.
3. Unity Catalog must be enabled. The workspace needs to be Unity Catalog-enabled, not just using the legacy Hive metastore.
4. Foundation model access via system.ai schema. You need access to foundation models through the system.ai catalog. Verify that the system.ai schema is accessible in your workspace by running a quick check in a notebook or SQL editor:
SELECT * FROM system.ai.served_models LIMIT 5;
If you get an error, your workspace may not have the system catalog configured properly. Your account admin can verify this is enabled.
5. Serverless budget policy with a nonzero allocation. You mentioned you have set the serverless budget quota, which is good. Just double-check that the policy attached to your workspace has a nonzero DBU allocation (not set to zero or unlimited-but-disabled).
6. Supported AWS region. Agent Bricks is currently available only in specific AWS regions (us-east-1 and us-west-2 at launch). If your workspace is deployed in a different region, the feature tiles will appear disabled regardless of other settings. You can check your workspace region in the account console under the Workspaces section.
7. Premium or Enterprise tier. You mentioned you have Premium, which should be sufficient.
MOST COMMON CAUSES
Based on what you described, the two most common reasons the feature remains disabled even after checking the above are:
- Region: If your workspace is not in a supported region, the tiles will stay grayed out. This is the most frequent cause.
- Specific preview toggle: There are multiple AI-related preview toggles. Make sure you are enabling the one specifically labeled "Mosaic AI Agent Bricks" (or similar naming), not just general AI assistant or Mosaic AI previews. The exact toggle name may vary slightly, so look for anything referencing "Agent Bricks" specifically.
NEXT STEPS
If all of the above checks pass and the feature is still disabled, I would recommend:
1. Try logging out and back in (or opening an incognito/private browser window) to refresh your session, as preview toggle changes sometimes require a fresh session.
2. If you are on Azure Databricks rather than AWS, note that Agent Bricks availability and region support may differ. Check the Azure-specific documentation for your region's support.
3. If everything checks out and the feature is still disabled, open a support ticket with Databricks. Include your workspace URL, region, and confirmation of all the prerequisite steps above. The support team can verify from the backend whether the feature entitlement is properly applied to your workspace.
HELPFUL DOCUMENTATION
- Agent Bricks overview and requirements: https://docs.databricks.com/aws/en/generative-ai/agent-bricks
- Knowledge Assistant setup guide: https://docs.databricks.com/aws/en/generative-ai/agent-bricks/knowledge-assistant
- Managing preview features: https://docs.databricks.com/aws/en/admin/workspace-settings/manage-previews
Hope this helps you get unblocked. Let us know how it goes.
* This reply used an agent system I built to research and draft this response based on the wide set of documentation I have available and previous memory. I personally review the draft for any obvious issues and for monitoring system reliability and update it when I detect any drift, but there is still a small chance that something is inaccurate, especially if you are experimenting with brand new features.
