Trust Center
Every rule we hold ourselves to.
REign AI is built on one principle: agents do the work, humans own the decisions. Here is exactly how that works in practice.
Governance
Agent autonomy levels
Every agent starts at Level 1. Promotion requires 4 weeks of 90%+ approval without edits. Any serious error drops it back immediately.
Level 1
Draft
Prepares work; a human approves before anything leaves the company.
Level 2
Act and report
Acts within set limits; a human reviews a daily digest.
Level 3
Autonomous
Acts alone on low-risk, reversible tasks with no human review required.
Anything involving money, contracts, health data, or publishing in a client's name stays at Level 1 permanently — no exceptions.
Operating rules
Operating rules every agent follows
Disclosure
Agents identify themselves as AI in every chat, email, and call. Several states require this. It also matches our brand promise.
Spending limits
Each agent has a monthly AI budget and no access to bank accounts. Agents can request payments but cannot send them.
Least access
Each agent connects only to the tools its job needs, with its own credentials stored in a secrets vault. No agent has more access than it requires.
Client data protection
No client data trains any model. We only work with AI providers whose terms allow business data and offer zero or limited retention.
Health data
We do not accept protected health information until HIPAA safeguards and signed business associate agreements are in place with every provider.
Artist consent
The Steward agent checks each work's consent settings before any use. No consent, no use — without exception.
Kill switch
Any agent, or all agents, can be paused instantly from one dashboard button. No escalation required.
Monthly ethics review
Every month, a sample of agent work is reviewed for bias, accuracy, and fairness to artists by our team and an independent advisor.
Human authority
Decisions that never go to an agent
A person remains legally responsible for REign AI. These decisions stay with humans, always.
- Signing contracts, artist agreements, and grant submissions
- Approving payments, refunds, payouts, and prices
- Hiring, firing, and anything about a person's employment
- Publishing in a client's name or on REign AI's public channels
- Security incidents, legal notices, and anything involving health information
Security
Security roadmap
We are building toward SOC 2 Type II certification. Here is where we are today.
- Per-agent credentials and secrets vault
- Full agent action logging and audit trail
- Human approval gate on all client-facing outputs
- SOC 2 evidence collection (Vanta / Drata)In progress
- SOC 2 Type II audit and certification
- HIPAA business associate agreements with all providers