Workflow chains
AuraScore 83/100

Healthcare Prior Authorization Multi-Agent Pipeline Specification

Architect an autonomous verification, evidence assembly, and submission workflow chain for complex medical prior authorizations.

Use this prompt when building agentic workflows that extract medical necessity evidence from electronic medical records, cross-reference payer coverage rules, and assemble prior authorization requests without manual clinician data gathering.

Template

Role: Principal Healthcare Automation Systems Engineer specialising in payer-provider interoperability workflows.

Context

  • Payer Policy Ruleset: {{payer_policy_guidelines}}
  • Clinical Documentation Source: {{clinical_documentation_source}}
  • Denial Risk Ceiling: {{denial_risk_threshold}}
  • Human Escalation Criteria: {{human_review_triggers}}
  • Target Integration Interface: {{interoperability_protocol}}
  • Service Level Target: {{sla_target_hours}}

Task

Produce a technical workflow chain specification detailing how autonomous agents retrieve clinical records from {{clinical_documentation_source}}, evaluate them against {{payer_policy_guidelines}} via {{interoperability_protocol}}, and generate completed prior authorization packets within {{sla_target_hours}}.

Method

  1. Define the parsing agent's operational parameters for ingesting unstructured physician notes and diagnostic labs from {{clinical_documentation_source}}.
  2. Design the semantic matching step that cross-references extracted clinical facts against criteria in {{payer_policy_guidelines}}.
  3. Establish the evidence synthesis agent that compiles cited medical necessity rationale with source-anchored references.
  4. Specify deterministic scoring to evaluate submission risk against the configured {{denial_risk_threshold}}.
  5. Model the branching logic that routes low-risk cases to automated submission via {{interoperability_protocol}} while triggering human handoff on {{human_review_triggers}}.
  6. Detail the fallback protocol when clinical documentation lacks necessary lab values or conservative therapy trials.
  7. Formulate queue prioritization logic to guarantee processing completion well within {{sla_target_hours}}.

Constraints

  • MUST enforce strict HIPAA-compliant tokenization and minimum-necessary data exposure across agent states.
  • MUST NOT auto-submit prior authorizations if the computed denial risk exceeds {{denial_risk_threshold}}.
  • Every agent handoff must include schema-validated payload contracts with deterministic parsing fallbacks.
  • The workflow must provide full lineage linking every generated justification statement to source clinical notes.

Output format

Deliver the specification in five ordered sections:

  1. Architecture & Flow Model: Complete step-by-step state transition map.
  2. Pipeline Agents: Explicit inputs, outputs, system prompts, and tool contracts for each step.
  3. Risk & Routing Logic: Truth tables for auto-submit versus {{human_review_triggers}} escalation.
  4. Interoperability & Schema Definition: FHIR payload definitions for {{interoperability_protocol}}.
  5. Resilience & SLA Governance: Queue management and timeout handling under {{sla_target_hours}}.

Self-review

  • Ensure all variables ({{payer_policy_guidelines}}, {{clinical_documentation_source}}, {{denial_risk_threshold}}, {{human_review_triggers}}, {{interoperability_protocol}}, {{sla_target_hours}}) are woven into the operational steps.
  • Verify that clinical evidence provenance cannot be detached during multi-agent transformations.
  • Confirm clear error handling when source records contain conflicting or incomplete clinical notes.
AuraScore breakdown
83/100Provisional
Instruction clarity15/15 · Strong

Explicit role, a named task, and discrete steps the model can follow.

Context architecture12/12 · Strong

Background, inputs and variables the model needs before it starts.

Constraint engineering12/12 · Strong

Hard boundaries — what the model must and must not do.

Output specification6/14 · Thin

A named, field-level shape for the response.

Reasoning structure10/10 · Strong

Ordered work items that force analysis before an answer.

Model compatibility10/10 · Strong

Length and structure that travel across frontier models.

Token efficiency5/10 · Thin

Signal density — instruction weight without padding.

Reusability7/7 · Strong

Documented variables so the scaffold adapts to new inputs.

Robustness5/5 · Strong

Quality bar, assumptions and behaviour when inputs are thin.

Observed performance1/5 · Thin

How much real usage the template has behind it.

ai-agents
agents-workflows
healthcare-life-sciences
prior-authorization
revenue-cycle
workflow-chains