How the Radiant Workflow Works
From the hospital imaging archive through on-site de-identification, blockchain-backed chain-of-custody, AI-ready dataset preparation, and proportional value distribution — every step governed, auditable, and controlled by the originating facility.
The Radiant Data Workflow
This is not just software — it is an operating system for governed healthcare AI data. Every step is documented, every movement is traceable, and every contributing facility receives proportional value.
🏥 Hospital / Imaging Center / Pathology Lab
Source systems remain inside the facility. DICOM studies, clinical imaging workflows, pathology data, and related operational systems are the starting point. PHI remains under facility control at all times.
🛰️ RAIHD Cadet Installed On-Site
Cadet deploys inside the hospital firewall. DICOM and related data are collected locally. PHI is removed inside the firewall — all 18 HIPAA Safe Harbor identifiers are scrubbed from headers. Burned-in pixel PHI is detected and permanently redacted. Questionable studies are quarantined for facility review.
✅ Facility Validation
The hospital reviews sample batches or validation reports produced by Cadet. Only approved de-identified data proceeds. The facility retains institutional control over which research categories are permitted — and can exclude any category that conflicts with community expectations.
🧭 RAIHD Commander Orchestration
Commander coordinates routing, migration, dataset preparation, chain-of-custody documentation, and operational monitoring. All data movement is structured, governed, and auditable from the central orchestration layer.
🔗 Blockchain Chain-of-Custody Layer
Dataset events are recorded on the blockchain layer — not PHI, but metadata: movement records, access events, de-identification events, dataset construction, and use records. This creates an immutable, auditable provenance trail.
🔬 Health AI Data Network
De-identified datasets are made available to qualified AI research partners through Health AI Data, Inc. — the partner-facing data network and research enablement entity. Researchers access governed, traceable, reproducible datasets from diverse patient populations.
💰 Revenue Share / Royalty Distribution
Contributing facilities receive proportional value from approved downstream usage. The automated royalty distribution system allocates revenue based on verified dataset contribution — creating a recurring data asset from imaging archives that would otherwise generate no value after the initial diagnostic read.
🔄 Future Return Workflow
In later phases, AI outputs — secondary captures, structured reports, segmentation objects, PDFs, or HL7/FHIR reports — can be routed back to the originating site when appropriate. The contributing facility not only provides data but can also benefit from the AI research it helps enable.
What Makes This Workflow Different
PHI Never Leaves the Facility
De-identification is completed entirely on-site, inside the hospital's own domain. This is not a cloud-first model. PHI stays where it belongs.
Every Step Is Auditable
Blockchain-backed chain-of-custody records document every movement, access, and use event. Hospitals and regulators can verify the full data lifecycle.
Facilities Receive Value
Contributing hospitals receive proportional revenue share — transforming dormant imaging archives into recurring strategic assets.
Who This Workflow Serves
Hospitals
Modernize archives, generate recurring revenue, and maintain institutional data control.
Imaging Centers
Improve workflow coordination and unlock research value from high-volume imaging data.
AI Researchers
Access governed, traceable, reproducible datasets from underrepresented patient populations.
Investors
See a defensible operating model — not just software — with patent-pending architecture and recurring value creation.
Common Questions About the Radiant Workflow
Does any PHI leave the hospital?
No. All de-identification happens on-site through RAIHD Cadet inside the facility's firewall. Only validated, de-identified data moves forward.
What is recorded on the blockchain?
Only metadata — movement records, access events, de-identification events, dataset construction, and use records. No PHI is stored on-chain.
How do hospitals control which research is permitted?
Each facility selects which research categories to approve or exclude at contract signing. The opt-out model ensures no facility is locked into categories it did not approve.
How does revenue share work?
Contributing facilities receive proportional revenue from approved downstream AI research usage, based on verified dataset contribution.
Ready to See How the Workflow Fits Your Environment?
The conversation starts with your environment — not a demo.