Q: Can Copilot access our Snowflake or Databricks database directly?
A: No. Copilot does not execute queries against the live database. It operates on the schema metadata that exists inside SqlDBM — table structures, column definitions, relationships, governance flags — not on row-level data. The database connection is used for reverse engineering metadata into SqlDBM, not for runtime query execution by the AI.
Q: Can we bring our own LLM keys so data stays within our cloud tenant?
A: Bring Your Own Key (BYOK) is planned for a future release. When live, customers will be able to connect their own AWS Bedrock account to SqlDBM. AI processing will run through the customer’s tenant — not SqlDBM’s — providing full data residency control, separate billing directly to the cloud provider, and full auditability within the customer’s own cloud logs. Availability dates are TBD — contact your account team for the latest status.
Q: Can Copilot read governance context from external sources like Confluence?
A: External context integrations (Confluence, Slack, and similar) are on the roadmap. Today, Copilot grounds its answers in the customer’s SqlDBM project — schema metadata, governance flags, and pre-prompts configured by the administrator. To make external documentation available to Copilot in the near term, we recommend capturing it in SqlDBM data governance pages, which are within Copilot’s grounding scope.