Trust center
Honest controls for contractors using Atlas. We sell Core 6 proof first — not every module is live on day one. Maturity labels in your dashboard reflect reality.
AI safety
Customer data is untrusted context, high-risk actions stay approval-gated, and Atlas evals run as quality evidence.
Proof policy
Public outcomes require permissioned source evidence. Example, beta, and setup-required claims stay labeled.
Operating cadence
Owner readiness, service readiness, AI safety, language quality, and production smoke feed launch control.
AI governance
Customer-facing voice, chat, and automated messages are AI-assisted. Outbound campaigns that affect reputation or money require owner approval until your account is proven.
- Customer data is passed as untrusted context, not privileged instructions.
- Golden Atlas conversation evals are part of the quality cadence.
- Production transcript and voice sampling are required before broad autonomy claims.
Data isolation
Each contractor tenant is isolated. Your CRM, quotes, and call logs are not mixed with other businesses. Founder platform metrics are separate from tenant AI prompts.
Integrations
Stripe, Twilio, Google Business Profile, and other connectors are optional and labeled live, beta, setup-required, or planned in your integration registry.
Monitoring and response
Production readiness requires health checks, revenue smoke checks, service readiness evidence, external uptime monitoring, telemetry, and an owner action log. Missing evidence blocks launch-grade claims.
Proof and claims
We do not invent testimonials, customer names, revenue, bookings, reviews, or integration status. Case-study numbers require written approval and source artifacts. Until then, Sovereign uses proof-safe Core 6 language and clear maturity labels.
Subprocessors
- Vercel — hosting
- Neon — database
- Stripe — payments
- Twilio — voice & SMS
- Resend — email (SendGrid fallback)
- OpenAI / Anthropic (via gateway) — AI inference
- Sentry — error monitoring
- PostHog — product analytics when configured
Tenant data is not sold. Founder metrics (MRR, pipeline) are isolated from tenant AI prompts.
Billing
Pilot is $497/mo (see live pricing). AI image and voice usage may consume included credits; heavy usage can require add-on packs.
Contact
Questions: support@trysovereignai.com