AI questions
AI
FAQ
How the AI works, what it sees, and what it never does.
Does the AI train on our data?
No. The AI agent uses your portfolio context inside your workspace only. Your data is never used to train shared models, and prompts and responses are not retained by the model provider.
What AI model does Vantage use?
Vantage uses Google Gemini (currently gemini-2.5-flash for live workloads) through the Lovable AI gateway. Model selection is documented and changeable per workspace.
How do you prevent the AI from hallucinating?
Answers are grounded in retrieval over your workspace knowledge base and structured data. The AI never invents IDs, owners, or dates; if context is missing, it asks. Every AI-generated brief is editable inline with a revision history.
Can I turn off the AI features?
Yes. Workspace owners can disable AI scoring, AI brief generation, and the knowledge base independently. The deterministic RICE/WSJF score continues to run when AI scoring is off.
Is there an audit trail for AI actions?
Yes. Every AI-driven score adjustment, brief generation, and knowledge-base query is logged in the immutable activity log alongside human actions.
Don't see your question? Browse all categories or contact the team.