Tone calibrated to your practice
The model is trained on your 50 most recent published replies and on your website copy, with a 2× weight for replies you kept as-is.
Feature · Voice Agent
AI replies in your practice's voice
The Voice Agent learns your tone from three sources: your previous published replies, your website's public copy, and a 30-second fine-tune you set on day one.
For reviews the Safety Agent authorises, the Voice Agent drafts a reply within 15 minutes. The draft is editable, versioned, and the final version is the one you approve — not the one the AI suggested.
What the Voice Agent does
Edit and regenerate
Every draft is editable. You can rewrite it, ask for a variant, or escalate the decision to another team member.
Versioning and audit
Every draft version is kept, with the author of the change and the timestamp. The audit log remains the source of truth for any reporting workflow.
Replies in multiple languages
The reply is drafted in the language of the review. Multilingual reviews receive a reply in each detected language.
How the Voice Agent drafts a reply
Once the Safety Agent authorises a review, the Voice Agent drafts a reply in the practice's tone, edits it, and hands it to the Patient Agent for publishing.
- 1
Check the Safety authorisation
The Voice Agent only drafts for reviews the Safety Agent has marked as cleared to publish. Sensitive reviews stay locked.
- 2
Read the tone and the context
The model pulls your 50 most recent validated replies and your website's home page to calibrate the tone.
- 3
Draft the reply
A draft is produced in under 30 seconds, with a length calibrated to the average of your previous replies.
- 4
Hand-off to the Patient Agent
The draft is stored, versioned, and handed to the Patient Agent for human validation and publishing.
Why now
Framework current as of September 2026.
Three bodies of law govern how you reply to an online review.
Revydo does not reinvent the framework. It absorbs it into every agent, so compliance does not depend on a forgetful assistant.
Legal framework
Consumer right to publish an online review
An online review is protected speech, not a marketing artefact. Hosting platforms and the reviewed professional both have obligations to the publisher, including the right of reply and the limits on removal.
EU Directive 2000/31/EC and national transpositions
Personal data
GDPR in force since May 2018
Every review contains personal data. The act of drafting a reply, storing it, or displaying it alongside the review is processing under GDPR. A careless reply is a separate data event, not a footnote.
Regulation (EU) 2016/679
Professional ethics
Codes of professional conduct
Where the practice is regulated (medicine, law, dentistry, physiotherapy, etc.), the professional code of ethics stays enforceable inside the reply. A draft that confirms a client relationship or reveals a covered fact is a disciplinary risk, not a marketing choice.
National codes of professional conduct
Frequently asked questions
Can the Voice Agent invent clinical facts?
No. The Voice Agent is trained to only produce relationship-, welcome- and operations-level sentences. Any mention of a treatment, diagnosis, or outcome is blocked by construction and escalated to the Safety Agent.
What if I don't like the draft?
You can rewrite it, ask for a new variant (keeping the same tone), or write the reply yourself. Every modification is versioned in the audit log.
How long does drafting take?
Under 30 seconds for an initial draft, under 15 minutes for an authorised review to receive a published reply (including human validation).
Does the model improve over time?
Yes. The Voice Agent re-trains its tone calibration with every validated reply, with a higher weight for replies you did not modify.
Go further
Discover review requests
The Voice Agent only drafts for replies the Safety Agent has authorised. For sensitive reviews, the alert reaches the practitioner with a measured draft.