Trust Center

Advice about your life has to be right. Here is exactly how we keep it grounded, governed, and private.

Grounded AI

Every personal fact — balances, APRs, goals, history — is read deterministically from your system of record and labeled authoritative. The model is instructed never to invent personal data.

Fail-Closed

If a fact is missing or unavailable, the assistant says so and offers to help you add it. It does not estimate, infer, or fill in from an example.

Personal GraphRAG

Your data lives in a private knowledge graph isolated per user, with row-level security across every store.

Two-Layer Separation

Central knowledge (shared methodology, compliance, safety) governs how we answer. Your personal data — never shared — determines what is true.

Governed Recommendations

Advice passes a constitutional + compliance layer before it reaches you; unsafe or non-compliant output is blocked.

Privacy by Design

Least-privilege access, encryption, and per-user isolation throughout. Plaid connectivity (full product) is read-only.

Looking for specifics?

AI Assistant

🤖

Welcome to AI Assistant

Loading AI agent...

Press Enter to send, Shift+Enter for new line

General information, not financial, tax, or legal advice.