Where Databricks data becomes action
Databricks holds the data and the building blocks. Sigma takes it the rest of the way, where business teams analyze, build apps and run agents under Unity Catalog.
Data to action
Business teams explore, plan and decide on live Delta tables in a spreadsheet, with no SQL or extracts.
Apps that write back
Inputs and approvals land in Delta tables, and business users change the app in a few clicks.
Agents that act
Business teams build agents that call your Genie Agents, then request approval and write back.
Lakehouse ROI
Bring the lakehouse to the business
Databricks gives data teams one place for data and AI. Business teams often still wait in a queue, then finish the work offline on extracts.
| Today | With Sigma |
|---|---|
| Backlogs. Extracts. Offline Excel. | Business teams self-serve on live Delta tables |
| Policies rebuilt in each BI layer | Unity Catalog policies apply at query time |
| Metrics that drift by team | Shared metric definitions in one data model |
| Insights handed off by email | Inputs and approvals written back to Databricks |
| App changes that wait on a ticket | Business users change apps in a few clicks |
One platform
Build the pieces in Databricks. Finish the work in Sigma.
Sigma runs live on the lakehouse you already govern. Business teams query, write back to Delta tables and call your Genie Agents, under Unity Catalog.
- Finance
- Operations
- Sales
- Marketing
- Supply chain
- Workbooks and dashboards
- Apps with writeback
- Sigma Assistant
- Sigma agents
- Live queries, no copies
- Writeback to Delta tables
- Assistant uses Genie Agents
- Agents call Genie Agents
- Databricks SQL
- Delta tables in your lakehouse
- Genie Agents
- Unity CatalogRoles and policies apply to each query
Finish the work where the data lives
Databricks provides the building blocks. Sigma is where business teams finish the work, with Unity Catalog governance built in.
Full detail
From the KPI to the row on live Delta tables
Teams drill from a KPI to the row across billions of rows, live on Delta tables. As a Lakehouse//RT launch partner, Sigma stays fast at high concurrency.
- No extracts or imports
- Drill from any chart to its rows
Trusted metrics
Define metrics once for people and agents
Build a Sigma data model on Unity Catalog tables, with relationships and metrics in one place. Dashboards, apps and agents answer from the same definitions.
- Relationships and metrics in one model
- Agents that answer from the model
Captured decisions
Write plans and approvals back to Databricks
Teams add forecasts, commits and notes beside live data in an input table. Each entry lands in a Delta table you control.
- Entries written to Delta tables
- Edit permissions set per table
- A record of each change
Finished work
Agents that take the next step
Business teams build Sigma agents in the workbook. They call your Genie Agents for context, then write changes back to Databricks once someone approves.
- Genie Agents available as tools
- Approval before anything is written back
- Updates posted to Slack
Inherited controls
Unity Catalog rules apply to each query
Queries run as each person with OAuth, or as a service account. Unity Catalog grants apply, and query history shows the SQL each element ran.
- OAuth run-as-user or a service account
- Row filters and column masks at query time
- Private connectivity on AWS and Azure
Teams that act on Databricks data with Sigma
How business teams put live lakehouse data to work

Weeks to days to ship new workflows. 14K+ r&d orgs on luma. 2M+ scientists using the data.Weeks to days to ship new workflowsHow Dotmatics ships new workflows in days with Sigma
100+ hours saved/month for planning teams. 80% of time reallocated to strategic, value-add work. 6 department use cases in year 1.100+ hours saved/month for planning teamsWhen this furniture company needed business intelligence, it found Sigma on Databricks
50% weekly analyst hours reclaimed from manual report building. 5 net-new use cases for portfolio projections and interactive reporting. $20B assets under management (aum) supported by sigma-governed data.50% weekly analyst hours reclaimed from manual report buildingMakena Capital Cuts Analyst Reporting Time by 50% and Builds a Firm-Wide Data Culture on SigmaGo deeper
For your Databricks rollout
The docs and guides admins ask for when connecting Sigma to Databricks.
Connect to Databricks
Set up the connection to a Databricks SQL warehouse.
Open the docsUnity Catalog in Sigma
How grants, row filters and column masks carry into Sigma.
Read the blogOAuth for Databricks
Run each query as the signed-in user with OAuth.
Read the blogUse Genie Agents in Sigma
Call your Genie Agents from Sigma Assistant and Sigma agents.
Open the docs
Databricks FAQ
What Databricks admins and security teams ask before they connect Sigma.
How does Unity Catalog governance carry into Sigma?
With OAuth, each query runs as the signed-in user, so their Unity Catalog grants, row filters and column masks apply. With a service account, map Sigma teams to Databricks permissions and add row-level security in Sigma.
Does Sigma copy our data out of Databricks?
No. Sigma compiles each request to SQL that runs on your Databricks SQL warehouse and keeps no copy of your tables. Input tables write to Delta tables you control.
Does each click run a Databricks SQL query?
No. Sorts and filters on loaded data run in the browser, and repeated queries can use Databricks SQL caching or a Sigma cached result. Logic many workbooks share can be materialized to a Delta table on a schedule.
Can Sigma use our Genie Agents?
Yes. Sigma Assistant and Sigma agents can call the Genie Agents you already maintain. Each Genie Agent runs in Databricks under the permissions of the person asking.
Where does writeback land, and is it audited?
Input tables write to Delta tables in a catalog and schema you choose. Each entry records who changed it and when, and edit permissions are set per table.
How do we keep traffic private?
Sigma supports AWS PrivateLink and Azure Private Link for Databricks connections, so traffic stays off the public internet. Reports for your security reviewers are in the Trust Center.
Can we trace a Sigma query in Databricks?
Yes. Query history in each workbook shows the generated SQL, where it ran and its timing, with the Databricks query ID to find it in Databricks.
See Sigma on your Databricks data
We’ll connect to your Databricks workspace and walk your team from data to action, with Unity Catalog in place.

