Rōvn · Investor Room
AI agent: checking…
All sections
Product

Facility Workflow Memo

Current truthRōvn master canon generation 8 · effective 2026-07-21. Earlier dated diligence documents are historical snapshots, not current deployment proof.Ask the canon-grounded agent →
AI Diligence Console

Facility Workflow Memo: Readiness and Operator

Status: July 2026 · aligned to canon generation 8
Investor read: the facility-side product, the human-control boundary, and what is demonstrable today

TL;DR

Facilities lose workforce capacity because hiring, credentialing, onboarding, privileging, payer enrollment, monitoring, and coverage operate as disconnected work. Rōvn's facility answer is two connected surfaces: Readiness, the paid wedge that shows what will block starts, billing, or coverage and routes the exact next action, and Operator, the complete facility product that runs the workflow end to end. The output of the loop is the Work Activation, approved to start, with proof: an organization-specific, evidence-backed state approved by a named human and bound to a signed receipt.

Doctrine: AI compresses the work. Source systems prove the facts. Humans make every credentialing, privileging, hiring, and clinical decision.

Readiness: the paid wedge

Readiness answers seven questions on the organization's own roster:

  1. Who is clear to start?
  2. Who is clear to practice?
  3. Who is clear to bill?
  4. Who is expiring?
  5. What is Rōvn working on?
  6. What requires a human decision?
  7. What future work is at risk?

The canonical dimensions are START, PRACTICE, and BILL. Readiness is deterministic and scoped: it never produces a blended worker score, it preserves ambiguity as explicit human-review states, and it pins every result to the exact evidence set, requirement version, and evaluation time. The published entry price is $2,500 per month; every tier above it is quoted.

Detection opens a Resolution Case

Readiness detects risk; it does not own remediation. Detection opens or updates a Resolution Case: one deduplicated owner of missing, expiring, ambiguous, or disputed work. One underlying issue creates one worker-facing ask, with tenant-isolated impact edges to every affected facility, file, payer, privilege, and future assignment. The facility sees three lanes:

  • Needs you: decisions, attestations, ambiguity, adverse facts, exhausted outreach.
  • Coverage at risk: unresolved cases threatening future work.
  • Rōvn working: grouped progress where staff have nothing to do.

There is no separate chase inbox. The action queue, work log, provider drawer, agent trace, worker task, and scheduling view are projections of the same case.

Operator: the complete facility product

Operator runs demand and requisitions, applicant intake, hiring workflow, readiness, credentialing, privileging, payer and billing readiness, onboarding, the worker file, Resolution Cases, monitoring, coverage risk, assignment eligibility, human decision queues, reports, integrations, and audit proof. It is the destination product, activated in stages, and it succeeds when coordinators manage exceptions and decisions instead of manually moving every file.

Human decision gates

GateHuman control required
Hire and offerNamed reviewer, decision, typed rationale, receipt
CredentialingCommittee or delegated reviewer, packet receipt, approval or deferral reason
PrivilegingPrivilege-specific reviewer or committee, privilege-to-competency evidence, approval or limitation reason
Assignment and coverageNamed human decides; Rōvn prepares version-bound actions, the organization's authorized system executes
Adverse actionNamed decision-maker, due-process path, reporting review where applicable

No AI-only hiring decision. No AI-only credentialing or privileging decision. No AI-only adverse action. The system may know an organization approved a person; it may never invent, transfer, or universalize that approval.

Policy is an explicit object

Every evaluation runs against a versioned RoleRequirement: organization, facility, role, specialty, service, location, engagement type, payer context, required evidence, freshness intervals, exceptions, approval authority, and effective dates. The same evidence can produce different readiness outcomes at different organizations. That is expected, not an inconsistency, and it is why local recognition survives portable evidence.

What is demonstrable in July 2026

  • Readiness demo: readiness.rovn.to, open with no login, restored 2026-07-22. Synthetic protagonist Maya Patel operating a 235-clinician roster at Sunbelt Surgical Partners.
  • Operator demo: operator.demo.rovn.to. Enterprise Operator surface for the synthetic Alder Crest Health Network, 1,248 synthetic workers.

Both demos run entirely on synthetic data: no PHI, no real roster, read-only demonstrations, never production customer deployments. No facility has yet completed the governed workflow on real data; that claim is reserved for a signed pilot with receipts and a named witness. Deeper credentialing, privileging committee, payer enrollment, and survey-export workflows are designed and being built in stages, and are described in 04.7 Regulated Workforce Expansion Plan.

Procurement-safe line

Rōvn Readiness is designed to show what will block starts, billing, or coverage and route the exact next action, so coordinators spend less time chasing and leaders see risk before it becomes lost capacity. Each facility applies its own requirements, and a named human makes every regulated decision.

Ask the AI agent about this section, the raise, compliance posture, or any cross-document question. Grounded in Rōvn canon generation 8, with on-page source citations.

Investor questions run through Google Cloud Vertex AI and are constrained to the hash-pinned Rōvn generation 8 canon. No PHI belongs in this room or its prompts.