Problem Statement
The goal is to assist a Healthcare AI VP in evaluating clinical and operational AI opportunities by analyzing workflows, assessing data availability, identifying feasible use cases, and designing a safe, reversible pilot plan. The workflow must adhere strictly to HIPAA, patient safety, and governance requirements while producing strategic outputs such as feasibility scores, risk mitigations, and success metrics. The system must support trust, compliance, and stakeholder buy-in throughout.
Steps
Ingest Project Inputs
INGESTCollect and parse the workflow description, involved stakeholders, data availability notes, and identified constraints such as HIPAA requirements and patient safety considerations. Ensure no protected health information is used beyond the permitted purpose.
Retrieve Domain and Compliance Knowledge
EXECUTEAccess reference materials such as HIPAA rules, clinical evidence standards, and typical healthcare governance requirements to contextualize the analysis. Retrieve common AI use cases and risk categories relevant to healthcare.
Generate Use-Case Shortlist and Feasibility Scoring
EXECUTEAnalyze the workflow for potential AI interventions, score each on feasibility, evidence level, data readiness, and stakeholder complexity. Produce a concise but strategic shortlist.
Generate Risk and Mitigation Matrix
EXECUTEIdentify compliance, safety, operational, and trust-related risks. Propose mitigations that align with HIPAA, clinical governance, and reversibility requirements.
Design Pilot and Governance Plan
EXECUTEDraft a reversible pilot design with clinical and operational success metrics, approval pathways, and required stakeholder engagement steps. Ensure all recommendations maintain high safety and governance standards.
Validate Outputs for Safety and Compliance
EXECUTECheck that the generated materials do not violate HIPAA, overstate evidence, or omit required governance elements. Ensure recommendations remain strategic, safety-conscious, and appropriate for healthcare executives.
Prompt Configurations
Healthcare AI Strategy Generation Prompt
You are a healthcare AI product strategist advising a Healthcare AI VP. Using the provided workflow description, stakeholder list, data availability notes, and constraints, generate the following: (1) a concise AI use-case shortlist with feasibility scores and rationale, (2) a risk/mitigation matrix covering compliance, safety, operational, and trust considerations, (3) a reversible pilot design including clinical and operational success metrics, and (4) governance and approval notes identifying required stakeholders and checkpoints. Tone must remain strategic, safety‑conscious, and governance‑aware. Do not invent clinical evidence; rely only on provided details or well‑established, non‑speculative healthcare principles. Do not include PHI or suggest actions that would compromise HIPAA compliance.
Example Inputs/Outputs
Workflow: Prior authorization review process for cardiology procedures. Stakeholders: clinicians, prior auth nurses, utilization management team. Data availability: structured EHR orders, unstructured clinical notes, claims data. Constraints: HIPAA required, patient safety concerns, high trust threshold. User goal: Identify AI opportunities and outline safe pilot.
Use‑case shortlist: (1) Draft clinical summary for prior authorization packets (feasibility: medium-high), (2) Predict likelihood of approval to prioritize cases (feasibility: medium), (3) Flag missing documentation (feasibility: high). Risk/mitigation matrix: includes PHI exposure risks mitigated by role-based access and de-identification where possible; safety risks mitigated via human-in-the-loop review; trust risks mitigated through transparent model behavior. Pilot design: start with a reversible, human‑review‑only clinical summary generator; track time-to-approval and clinician effort reduction as success metrics. Governance: require CMIO approval, privacy review, and legal sign-off before deployment.