Security and compliance
Outbound touches other people's inboxes and phone lines. That makes compliance an engineering problem, so we built it as one. Here is exactly what runs and what does not exist yet.
Before every action
Not a nightly job, not a review queue. A function that runs before every single email and every single call, and records its verdict either way.
Suppression check
Global and account level, inside the same transaction as the send
Voice consent
Published business lines by default. Mobile numbers require recorded consent
Calling hours
Computed from the prospect's own timezone, not yours
Attempt caps
Four call attempts maximum, spread across different parts of the day
AI disclosure
Stated in the opening line of every call, every time
Opt out
Honored immediately, permanently, and across every channel
On AI voice specifically. The FCC has held that calls using AI generated voices fall under the same consent rules as prerecorded robocalls. Revio therefore targets published business lines by default. Calling a mobile number requires that you record prior express written consent for that contact, with a source and a timestamp, and the system will not place the call without it. We build no feature whose purpose is evading spam filters, caller identification, or consent requirements.
Platform controls
Every table carries a tenant id and a Postgres row level security policy. Isolation is enforced by the database, not by application code remembering a WHERE clause. A query without the right session context returns nothing at all.
TLS 1.2 or higher in transit. AES 256 at rest. Mailbox credentials and OAuth tokens are stored with envelope encryption and never written to application logs.
Four roles: owner, admin, member, viewer. Enforced at the API boundary and again at the row level. Enterprise accounts add SSO and SCIM provisioning.
Every send, call, configuration change, and export writes an immutable audit record with actor, target, before, after, and timestamp. Records cannot be updated or deleted, including by us.
Call recordings and transcripts are encrypted at rest with a retention window you set. Delete a prospect and the associated media goes with it.
Inbound webhooks are HMAC verified with a replay window. Per tenant and per endpoint rate limits are applied at the edge.
Current status
Security pages usually list aspirations as if they were facts. This one separates them, because you are going to ask in the security review anyway.
That is a reasonable order to do it in. Send us your questionnaire and we will answer it directly rather than pointing you at a trust badge.