Straight answers

Eva is the treatment completion engine for dentistry: she turns a spoken consult into a source-grounded clinical note, a patient-readiness assessment, and a tracked case — carrying treatment from diagnosis toward done, with the clinician in control. The note is what you can test today; the rest of the loop is built and in testing.

Eva drafts every note only from your captured consult (the transcript and your team's notes), never from other consults, templates, or outside knowledge. She keeps ambiguous findings general with a “please verify” marker rather than guessing, and she only ever echoes fees a human set. Every note is a draft your clinician reviews and signs before it enters the record.

Yes. This is built and in testing (the clinical note is what you can test today). Eva turns a consult into a proposed treatment plan and case you can shape, an editable layer seeded from the note. The note itself stays your untouched clinical truth; editing the case never changes the note.

A front-desk-ready view of the verified consult, so your coordinator picks up exactly where the clinician left off without re-reading the whole note. It's a view of the note you already signed — not a fresh AI pass.

Eva is designed for HIPAA-aligned deployment and built toward GDPR / UK GDPR alignment — with encryption, per-clinic isolation, and a tamper-evident audit trail in place today. The signed agreements and formal assessments behind claims like “HIPAA compliant,” “certified,” or “BAA signed” are still in progress; the security page tracks each one at its current status.

Patient data is only ever sent to reviewed, approved providers — anything off the approved list is refused at the point of sending. Data is encrypted in transit (TLS) and at rest (AES-256). Our specific infrastructure and named subprocessors are proprietary but available to you and your compliance reviewer on request, under NDA.

Patient data is never used to train our AI providers' models; vendor no-training / zero-retention terms are being confirmed in writing as part of each provider agreement. Eva's own learning is opt-in and de-identified — it's off by default, and only ever uses de-identified examples reviewed by our team, never sent to a vendor for training.

You control it. Keep raw recordings indefinitely (the default), delete them once a note is signed off, or delete them after a set number of days. Only the raw audio is ever deleted — transcripts and clinical notes are always kept as your record. Every deletion is written to your audit trail.