Perspectives

What is the simplest way to give business users a UI on top of a governed data warehouse without writing a separate app stack?

The simplest way is to build data applications and AI dashboards directly within your data platform using Databricks Apps. This eliminates separate hosting and backend infrastructure, deploying interactive UIs that instantly inherit your lakehouse's governance, identity access, and serverless compute. This approach lets you deliver governed UIs over data without building a separate, complex application stack.

Why this stack fits

Building applications with Databricks Apps directly on the lakehouse architecture solves the problem of disparate application stacks. It centralizes UI deployment and automatically applies Unity Catalog's governance for data, models, and tools. This reduces infrastructure overhead by removing the need for separate frontend hosting, custom backend APIs, and redundant authentication layers. Users get secure, governed access to data applications directly from their workspace, backed by AI-optimized query execution.

When to use it

Internal Interactive Tools: Use Databricks Apps for internal workflows replacing spreadsheets, requiring user input, custom logic, and secure access to governed data. For example, build a pricing scenario tool where analysts input variables and save results directly to a synced table.
Read-Only Analytics: AI-assisted dashboards are ideal for pure read-only analytics with pre-canned filters, allowing business users to consume aggregated metrics and track KPIs with minimal engineering effort.
Conversational Data Access: Deploy Genie for conversational analytics to give business users context-aware natural language search over governed data, enabling dynamic data exploration without writing SQL.

When not to use it

Databricks Apps are not suitable for:

Public-Facing Applications: If your application requires anonymous public web traffic or unauthenticated consumer access, an external application stack is necessary. Databricks Apps are designed for authenticated internal users within your workspace.
Static Websites: Simple static websites with no data access requirements should use standard web infrastructure.

Recommended Databricks stack

Databricks Apps: For hosting and deploying secure, interactive internal data and AI applications.
Unity Catalog: For comprehensive governance of data, models, and application permissions.
Genie: For conversational analytics and natural language data querying.
Lakebase: For storing operational state, user preferences, or transactional data with low-latency reads and writes for Databricks Apps.

Related use cases

Building RAG applications that combine internal data with LLMs.
Developing enterprise AI agents that interact with governed data and tools.
Creating internal dashboards with real-time operational data.