Your data never leaves your control.
AI in insurance only works if you can defend the answer and the data behind it. Here is how we make sure you can.
Tenant isolation by default
Every row in our database is tagged with your agency's tenant ID. Postgres Row Level Security blocks cross-tenant reads at the database layer, not just the app layer. Your data cannot leak to another agency, ever.
Citations on every answer
No answer ships without clickable citations to the source document and page. If you cannot point to where it came from, we did not say it.
PII redaction in chat logs
SSN, credit card, DOB, EIN, and phone numbers are masked before chat history is persisted. The model sees raw text in flight to give a useful answer, but storage is redacted by default. Org admins can switch to full or none.
Audit trail
Every chat, every proposal, every certificate has a full audit trail. Hand it to your E&O carrier without flinching.
Frequently asked
Where is my data hosted?
Encrypted at rest in a managed Postgres cluster, with backups in the same region. Data never leaves the platform without an explicit export.
Do you train models on my data?
No. Your documents and conversations are not used to train any base model. They power retrieval for your tenant only.
What about HIPAA and external audits?
We follow industry-standard controls today: tenant-isolated Row Level Security, encryption in transit and at rest, PII redaction by default, and a full audit log. HIPAA-aligned deployments are available for Network and Enterprise contracts. Formal third-party attestation is on our roadmap.
Can I delete my data?
Yes. Org admins can delete documents, conversations, and the entire tenant on request.