Skip to main content
The Elementary AI Agent automates the workflows that slow data teams down most: finding coverage gaps, triaging incidents, fixing broken pipelines, enriching metadata, and optimizing query performance. It’s one agent with a set of capabilities, each built for a specific workflow. You work with the same agent whether you open it from an incident, the catalog, the test coverage screen, or a Slack thread. Every action the agent takes goes through a pull request or requires explicit confirmation. Your team stays in control of what gets merged.

​
Capabilities

AI Governance

Enriches metadata and enforces documentation standards across your data assets.

AI Triage & Resolution

Investigates incidents, traces root causes, and opens pull requests with proposed fixes.

AI Test Coverage

Identifies coverage gaps and recommends or generates tests for the right assets.

AI Catalog

Finds the right data assets in natural language, with their relationships, usage, and health.

AI Performance & Cost

Detects slow or expensive models, flags redundant tests that slow down the pipeline, and suggests optimizations.

AI Code Review

Reviews every pull request that touches your dbt models, with a summary comment and findings on the changed lines.

​
Where you can use it

Slack agent

Tag @Elementary in an alert thread to investigate without leaving Slack.

MCP connections

Connect the tools the agent can reach, and reach Elementary from your own tools.

​
What the agent works from

The agent reasons over what Elementary already collects from your connected integrations, tied together through lineage:
  • Lineage across your warehouse, dbt models, BI dashboards, and code, down to the column level
  • Test results and incidents from dbt, Python, and cloud tests, including history and coverage gaps
  • Run and performance history for models, tests, and orchestration jobs, including execution time and cost
  • Code changes from your connected repository, such as recent commits and pull requests
  • Metadata such as descriptions, tags, owners, and table usage
This is what lets the agent trace a failure back to the commit that caused it, or tell you which dashboards an incident affects, without you assembling the picture first.
The more integrations you connect, the more the agent can see. Lineage between dbt and BI tools, for example, only works when both are connected. See integrations.

​
How it works

The agent runs on self-hosted models through Amazon Bedrock. No data leaves your environment, and all AI features are opt-in by default. Agent behavior is governed by AI policies you define: reviewable, adjustable, and aligned to your governance standards.

AI credits

What uses AI credits, and how to get more from them.

AI Privacy Policy

How AI features are secured and what data is accessed.
Want to learn more? Our experts will be happy to chat.