Q: What can Copilot actually do today?
A: Copilot covers five capability areas:
- Reverse engineering from unstructured inputs — PDM files, DDL, Excel, CSV, PDF, Word, images, or natural language requirements can be converted into a working data model.
- Talk to your model — ask any question about an existing project in natural language and get answers grounded in the full schema context.
- Bulk governance — flag PII, apply naming conventions, and add descriptions across many objects via a single prompt.
- Role-based responses — responses are calibrated to whether the user identifies as an architect, engineer, governance lead, or consumer.
- Pre-prompts — organizational standards taught to Copilot once are then applied to every interaction across every team member.
Q: What file formats can Copilot accept as input for reverse engineering?
A: PDF, Word, Excel, CSV, plain text, images (PNG/JPG), Draw.io diagrams, and DDL files, as well as freeform text pasted directly into the prompt field. Large Excel files work best when saved as CSV — Excel files include zipped internal content that can increase file size significantly without adding useful information.
Q: Can Copilot generate a logical model from existing DDLs across source systems?
A: Yes. Load DDLs from source systems into SqlDBM and Copilot will generate a draft logical model or mapping. Standard physical-to-logical conversion is also available as a non-AI pathway for teams that prefer a deterministic workflow.
Q: Can Copilot auto-create Data Vault structures?
A: Yes. Given staging tables and guidance in the prompt or pre-prompt, Copilot generates Data Vault structures — hubs, links, and satellites — aligned to Data Vault conventions. Results are improved by supplying project-level context through pre-prompts and standards.
Q: Does Copilot respect our naming conventions, templates, and governance rules?
A: Yes. If templates, global modeling standards, or pre-prompts are configured at the project level, Copilot uses them across every interaction. This is how organizations enforce consistency without reviewing every AI output by hand.
Q: Can Copilot work across multiple projects?
A: Not today. Copilot’s context scope is the current project. Cross-project governance validation is planned. Until it ships, global modeling — which lets you reference objects from other projects — is the recommended pattern for cross-project consistency.
Q: Can Copilot access version history — who made what change and when?
A: Version history context in Copilot is on the roadmap. Today, Copilot operates on the current state of the project. In the meantime, the platform’s Compare Revisions view lets users diff any two versions and see field-level and table-level changes. Revision flags (development, released, hotfix, etc.) can be used to mark meaningful milestones.
Q: Can Copilot suggest improvements, best practices, or impact analysis?
A: Yes. Copilot can review the current model and recommend schema improvements, normalization opportunities, or impact analysis for a proposed change. Because it has full project context, it can identify whether a proposed entity already exists in the model and recommend reuse rather than duplication.