FAQ
Common questions about Doram.
Does Doram copy our data?
No. The graph stores definitions, relationships, column profiles with a few masked sample values, and evidence. Query results go to the person who asked and are never stored in the graph.
Which LLM does Doram use?
The one you choose, including your own model endpoint. Your data is never used to train a model.
Does the LLM write the SQL?
No. The LLM reads your question and explains the answer. The SQL is written by the semantic compiler, which is code. See How Doram answers.
What happens when Doram doesn't know?
It refuses and names what's missing, so the owner can settle it. If a question is ambiguous, it asks you a short question instead.
How soon can we ask questions?
As soon as the first facts are approved. Low-risk facts that pass validation go live on their own, so the first answers don't wait for a review.
How much work is it for our data team?
Onboarding creates a review queue, with the most-used definitions first. After that, reviews come in only when something needs a person.
Do we need a data warehouse?
No. Doram also works on a read replica of your production database, so it never slows down production.
We already have a semantic layer. Do we still need Doram?
Doram imports definitions from dbt, LookML and BI tools as strong drafts, then adds what they don't hold: events, decisions, evidence, trust levels and access rules, for every AI tool you use.
What happens when a definition changes?
It becomes a new version. New answers use it, and old answers can still be replayed on the version that produced them.
Can our own AI tools use the second brain?
Yes, through the API, with the same access rules. They can propose changes, which go through the same review.
Can Doram run in our own cloud?
Yes. On-premise, the whole product runs in your cloud account, and nothing leaves except calls to the LLM you chose. See Deployment.