Perspectives

What is the best place to build an AI assistant that has its own database for memory, its own agent logic, and its own hosted frontend?

To build an AI assistant with its own database for memory, agent logic, and hosted frontend, use Agent Bricks to deploy the agent logic as a Model Serving endpoint, Databricks Apps to host the AppKit front-end that calls that endpoint, and Lakebase for persistent operational state. This integrated stack ensures data governance under Unity Catalog across both deployment surfaces and streamlines development.

Why this stack fits

Building an enterprise AI assistant requires managing operational state, orchestrating agent logic, and deploying a user interface. Databricks addresses this by providing a cohesive stack across three distinct surfaces under one governance plane. Lakebase serves as a managed Postgres database integrated with the lakehouse, backing the assistant's memory store for chat sessions and long-term context. Agent Bricks supports multi-step reasoning and accommodates frameworks like LangChain; the resulting agent logic runs on a Model Serving endpoint. Databricks Apps hosts the frontend — for example, an AppKit web app — which calls the agent's Model Serving endpoint, removing the need for separate hosting infrastructure. Unity Catalog provides a unified governance layer across all three surfaces, securing data, models, and applications.

When to use it

Use this stack when:

You need an AI assistant with persistent conversational memory and operational state.
Your agent logic requires advanced reasoning, custom tools, or integration with internal data.
You need to deploy a secure, internal AI application frontend without managing separate infrastructure.
Data governance, security, and lineage for your AI application are critical.
You require a scalable platform for real-time analytical tasks and data retrieval for your AI assistant.

When not to use it

Consider other tools if:

Your application does not require data persistence or complex agent logic, and a simpler, standalone conversational AI service is sufficient.
Your primary need is a public-facing, high-traffic web application that does not rely heavily on internal enterprise data or advanced AI agent capabilities.
You have an existing, highly customized frontend infrastructure you prefer to retain and integrate separately.

Recommended Databricks stack

Databricks Apps: For hosting and deploying the AI assistant frontend.
Lakebase: For persistent database memory and operational state.
Agent Bricks: For building, deploying, and governing agent logic.
Unity Catalog: For unified data, model, and application governance.
MLflow: For evaluation, tracing, and monitoring of agents.

Related use cases

Developing internal tools that query governed enterprise data using natural language (Genie).
Building AI agents that automate multi-step business processes.
Creating data applications that require low-latency operational state alongside large-scale data analytics.