Databricks Agent Bricks

Tracked · below panel floor — earned mentions are accruing toward the next scored window.

Measures earned, engagement-weighted share of voice across the GTM voices panel. Methodology →

Agent Bricks occupies the agent-building layer inside Databricks' platform: instead of asking teams to build agents from scratch, it auto-generates and optimizes agents against a company's own lakehouse data, using the same governance and model-neutral serving that underpins the rest of the Databricks stack. As a specific product line inside a much larger platform story, it is still early in the broader conversation, surfacing as one piece of the wider Databricks agent build-out rather than as a topic of its own so far.

Measures how easily your agents can build on it — API, MCP, CLI, SDK, docs depth. Methodology →

Agent Bricks bundles agent-building, evaluation, and governance into one surface: no-code builders and IDE-based development for multiple frameworks, continuous evaluation with LLM-as-judge scoring, and governance controls (role-based access, rate limits, prompt-injection defenses) applied uniformly across models. It connects to OpenAI, Anthropic, Google, Meta, and open-source models through a single contract, integrates MCP for external tool access, and deploys agents as serverless REST services with persistent memory through Lakebase.