A Framework for AI Care Orchestrators, Nurse Stewardship, and the Healthcare AI Operating Layer
Healthcare does not merely need smarter tools. It needs a new layer of stewarded intelligence that can hold together care, workflow, governance, and human values.
Pre-Directive evidence: This working paper predates NIN–NAIO Master Directive v1.1 and retains period terminology. It is preserved for provenance, not presented as current architecture, entity status, governance authority, clinical readiness, or conformance.
Healthcare is not transformed by isolated artificial-intelligence features. It is transformed by a governed orchestration layer that coordinates many kinds of intelligence — across models, workflows, people, and settings — into coherent, accountable, and humane action.
This paper names that orchestration capacity a distinct discipline, Care Intelligence, and argues that the profession already operating inside the orchestration loop of care — assessing, coordinating, escalating, monitoring, documenting, and sustaining continuity — is nursing. Building on established frameworks for care coordination (AHRQ), trustworthy and socio-technical AI (NIST), and the ethics and governance of AI for health (WHO), and on lifecycle controls for AI-enabled medical devices (FDA, IMDRF), it defines Care Intelligence operationally and sets out the mindset, the architecture, the administrative applications, the system configurations, and a staged roadmap by which institutions can build a trustworthy Care-Intelligence layer with nurses as its stewards.
It is deliberately confident about direction and design principles, and measured about enterprise return-on-investment claims, which require institution-specific evidence.
Durable transformation comes from a governed orchestration layer that composes many intelligences — not from isolated AI tools that shift the burden of integration onto clinicians.
The deeper risk is not a wrong answer but a right answer to the wrong objective — a metric optimized while the whole degrades. Govern objectives, not only outputs.
Nursing already runs the coordination loop, and the clinical reasoning cycle every nurse learns — assess, plan, act, evaluate — is structurally the feedback loop modern agentic AI is built on.
The orchestrator's competence — clinical, systems, governance, metacognitive, and communication — organized as a competency ladder outside parties can inspect.
A unified, sovereign orchestration layer — a governance-and-execution fabric across models and workflows — not a single model, and not a regulatory category.
Start with one auditable use case, govern before scaling, layer the intelligence, and treat deployment as a continuous learning loop — consistent with NIST, WHO, FDA, and IMDRF.
The public distrusts healthcare AI and trusts nurses more than any other profession. That gap is the opportunity — nurses are the credibility bridge healthcare AI is missing.
Operating commitments meant to shape institutional culture before they shape technology.
Preserved as a June 2026 snapshot. Its period terminology and project descriptions are not current Directive v1.1 authority or conformance claims.
Domondon, R. (2026). Care Intelligence: A Framework for AI Care Orchestrators, Nurse Stewardship, and the Healthcare AI Operating Layer (Version 1.0). Robert Domondon / NAIO project / Nurse Intelligence Network project. https://nurse-ai-os.org/care-intelligence/
Robert Domondon writes from a career that has spanned nearly every floor of the health system — bedside nursing, medicine, hospital administration, corporate strategy, public health, and healthcare communication. He is the founder and steward of the NAIO and Nurse Intelligence Network project initiatives, part of the Florence Media Network.
The project work advances a single conviction: that as intelligence becomes abundant, the judgment that governs it must become more capable — and that nurses, the most trusted profession and the natural stewards of the care process, are the ones to lead it.
This is a working white paper published openly for comment and is not clinical, legal, or financial advice. Coined terms (“Care Intelligence,” “AI OS”) are defined in the paper and tethered to existing governance frameworks. Selected return-on-investment and market-size claims are kept deliberately measured pending institution-specific evidence.