Hub

Workflow and task execution engine

SCRIMED turns healthcare workflows into sandboxed, audited, human-reviewed task plans.

CarePath AI, DocuTwin, TrialCore, Sanar AI, PayerIQ, and future services map to synthetic fixtures, sandbox runtimes, TrustQA, human approvals, result validation, and production deny gates.

Statussynthetic-ready
Workflows3
Ready3
Attention0

Agent Sandbox Runtime

Every agent receives isolated memory, files, workflow tools, and audit logs.

SCRIMED Atlas Intelligence Core v1 is a synthetic pilot and enterprise assessment operating layer. Medical-record, claims, payer, wearable, and telehealth workflows remain gated until tenant approval, BAA readiness, live connector controls, durable audit, and human review are approved.

Prior Authorization Agent

Prior authorization support

Per-task sandbox with synthetic policy, packet draft files, scoped memory, and review-only tools.

No payer submission, coverage guarantee, or medical-necessity determination.
  • Memory: session, operational, knowledge
  • Tools: policy evidence retrieval, packet draft builder, missing-evidence detector
  • Audit: task opened, policy cited, packet drafted, human review requested
Scheduling Agent

Referral intake automation

Referral-specific sandbox with routing rules, missing-information checklist, and escalation policy.

No autonomous referral acceptance, clinical triage replacement, or patient outreach.
  • Memory: session, operational
  • Tools: referral parser, routing constraint map, review queue generator
  • Audit: referral structured, gap detected, routing recommendation drafted, escalation created
Revenue Cycle Agent

Claims and denial risk review

RCM sandbox with claim metadata fixtures, policy references, and reviewer disposition state.

No final billing, coding, claim submission, or reimbursement guarantee.
  • Memory: session, operational, knowledge
  • Tools: denial-risk detector, appeal outline builder, documentation gap mapper
  • Audit: claim reviewed, risk signal surfaced, appeal draft prepared, review disposition recorded
Documentation Agent

Ambient documentation review

Documentation sandbox with draft-only note files, source trace, and signature-disabled output.

No final note, record filing, diagnosis insertion, or clinician-signature replacement.
  • Memory: session, operational
  • Tools: draft note builder, source trace mapper, missing context prompt generator
  • Audit: draft generated, source trace attached, review checklist created, signature blocked

Task execution engine

Task plans route through Planner, Router, specialist agents, TrustQA, and approval checkpoints.

enterprise-assessment, synthetic-pilot

Enterprise Workflow Assessment

Planner Agent. classify workflow, map value signal, attach governance requirements, select evidence needs

Interoperability Agent, Governance Agent, TrustQA Verification Agent
  • Approvals: Executive sponsor review, Governance review
  • Blocked: live connector execution, patient outreach, payer submission
synthetic-pilot, enterprise-assessment

Prior Authorization Packet Plan

Prior Authorization Agent. identify payer policy source, map missing evidence, draft review packet, attach reviewer checkpoint

Prior Authorization Agent, Revenue Cycle Agent, TrustQA Verification Agent
  • Approvals: Payer or RCM review, Governance review
  • Blocked: coverage guarantee, medical necessity determination, autonomous payer submission
enterprise-assessment, synthetic-pilot

ACCESS-Aligned Monitoring Plan

Sanar AI. map chronic-care track, identify wearable/telehealth signal, define outcome reporting, attach clinician-guided review

Clinical Intelligence Agent, Interoperability Agent, Governance Agent
  • Approvals: Clinical review, Executive sponsor review, Governance review
  • Blocked: reimbursement guarantee, autonomous patient monitoring, clinical outcome claim