Conversational analytics
Which platforms fit, and when we're not the right choice
Mark Your Data Assistant is for teams whose definitions live in dbt, with BigQuery as the warehouse today. If your definitions live in a platform's own semantic model and you don't use dbt, that platform's assistant is the simpler choice.
How it works out of the box
What it works with today. BigQuery with a read-only account, a dbt semantic layer, and Microsoft Teams or Slack.
Questions to ask any vendor. Can you trace every answer to a definition or SQL? Is there a feedback loop, and where does a fix land? Can all business users ask questions? Does it use your dbt definitions?
Where platform assistants win. They're built into your platform, so there's no new vendor to approve, and the platform enforces access per user.
Other platforms. On Snowflake, Databricks or another chat tool? Tell us, because we build the next connection with a design partner.
What it means for you
- Manager: You can compare tools on questions that stay relevant.
- Analytics engineer: If your definitions live in dbt, they stay in dbt.
Questions and answers
Which warehouses does Mark Your Data Assistant support today?
Mark Your Data Assistant supports BigQuery today, through a read-only account. Teams on Snowflake, Databricks or another platform can tell us what they use, because we build the next connection with a design partner. Support for other warehouses therefore follows a real project rather than a published list of connectors.
We use Databricks: should we use Genie instead?
Genie is the simpler choice if your definitions live in Databricks and you don't use dbt, because Genie is built into that platform. Mark Your Data Assistant answers from dbt definitions and doesn't support Databricks yet. If your definitions live in dbt, a design partnership is the way in.
We use Snowflake: when is Cortex Analyst the better choice?
Cortex Analyst is the simpler choice if your definitions live in Snowflake and you don't use dbt, because Cortex Analyst is built into that platform. Mark Your Data Assistant answers from dbt definitions and doesn't support Snowflake yet. If your definitions live in dbt, a design partnership is the way in.
How is Mark Your Data Assistant different from Copilot in Power BI?
The difference is where your definitions live. Copilot in Power BI works from Power BI semantic models, so it suits teams whose definitions live there. Mark Your Data Assistant answers from your dbt semantic layer, labels each answer governed or unverified, and shows your data team the gaps to fix in dbt.
What happens when you add support for our platform?
We build the connection with you as a design partner. Mark Your Data Assistant keeps the same labels and SQL guard: a new warehouse is a new connection behind them, and another semantic layer gets its own adapter. Your definitions stay where they are, and the assistant reads them from there.