Governance
Governance is the product.
NAIO governance and EDENA risk tiers are not compliance overhead. They are the reason regulated health systems and institutions will trust the program. Governance is built before the agents run. Never retroactively.
NAIO Institute
The Governance Institution
NAIO retains oversight authority over agent registration, risk tier assignment, budget governance, data zoning enforcement, and audit review cycles. Every agent, every workflow, every deployment answers to NAIO.
Agent Registration
Every agent must be registered before deployment. Required: identity, role, governance scope, risk tier, lifecycle, and human steward.
Stewardship
Each agent has a named human steward responsible for monitoring performance, reviewing escalations, and ensuring governance compliance.
Certification
Three credential levels: NAIO-C (Certified), NAIO-A (Advanced), and NAIO-F (Fellow). Each level unlocks greater governance authority.
Audit & Review
Every agent undergoes governance review every 30 days or upon escalation trigger. Immutable audit logs with defined retention policies.
EDENA Risk Classification
All AI actions must be classified before execution.
EDENA is the ethical decision engine embedded in Florence operations. Its core rule is simple and absolute: classify first, execute second. No exceptions.
Green Zone
Low-Risk / Mechanical
Informational tasks, research, drafting, market analysis, content outlines, and cohort materials. Fully automated with periodic review.
Oversight: Periodic review
Yellow Zone
Operational / Supervised
Actions affecting systems or stakeholders. External outreach, content publishing, pricing modifications, financial projections, and entity registration.
Oversight: Approval queue required
Red Zone
High-Risk / Human-Only
Irreversible decisions. Clinical actions, legal commitments, production deployments with PHI, financial commitments, and NAIO governance policy changes.
Oversight: Mandatory human authorization
The Agentic Loop
The standard loop pattern.
An agentic loop is a governed workflow cycle in which agents, skills, tools, memory systems, and human stewards repeatedly transform signals into action and action into improved institutional intelligence.
A loop without governance gates, monitoring, and feedback is not an agentic loop under this charter. It is merely an automation.
Escalation Protocol
Four levels of escalation.
No Red-tier decision may be executed autonomously. The escalation path ensures that risk is contained, investigated, and resolved at the appropriate level of human authority.
Automated Monitoring
Continuous system-level observation of agent behavior, performance metrics, and anomaly detection.
Steward Review
Named human steward reviews flagged items, approves threshold breaches, and adjusts agent parameters.
Governance Escalation
NAIO governance body convenes for policy-level decisions, risk tier reclassification, or systemic issues.
System Intervention
Emergency suspension, agent retirement, or infrastructure-level changes to contain risk.
Auditability
Every loop produces a verifiable audit record.
Logs must be immutable, retained according to policy, and available for operational, governance, security, and compliance review. The audit log is a product feature, not a burden.
Required Audit Fields
Governance is not constraint. It is structural integrity.
Institutional buyers trust NAIO-governed ventures over non-credentialed startups. The audit log, the risk tiers, the human stewardship — these are what make the Florence Ecosystem trustworthy at institutional scale.
Talk to Us About Governance