FlashAgents Orchestration

Synthetic data only

FlashAgents orchestration

1User question
2Orchestrator
3Specialized agents
4Evidence collection
5Validation
6Human review
7Final answer

Orchestrator Agent

Available in Demo

Plans the task and routes to specialists.

Inputs: User question, context
Outputs: Execution plan
Permissions: Read-only routing

Clinical Intelligence Agent

Available in Demo

Summarizes chart evidence.

Inputs: Diagnoses, labs, notes
Outputs: Evidence-linked summary
Permissions: Read clinical data

Risk Adjustment Agent

Available in Demo

Compares current vs projected HCCs.

Inputs: Diagnoses, claims
Outputs: Suspected HCCs for review
Permissions: Read; no coding writes

Claims Analytics Agent

Available in Demo

Aggregates and ranks spend.

Inputs: Claims lines
Outputs: Cost tables, trends
Permissions: Read claims

Population Health Agent

Available in Demo

Builds cohorts and gap lists.

Inputs: Members, gaps
Outputs: Prioritized cohorts
Permissions: Read

Financial Intelligence Agent

Prototype

Models financial scenarios.

Inputs: Cost, RAF, assumptions
Outputs: Scenario ranges
Permissions: Read

Medical Research Agent

Planned

Retrieves guideline context.

Inputs: Clinical question
Outputs: Cited references
Permissions: Approved sources only

Quality Management Agent

Prototype

Maps gaps to quality measures.

Inputs: Gaps, encounters
Outputs: Measure views
Permissions: Read

Care Coordination Agent

Prototype

Drafts outreach tasks.

Inputs: Queue, member
Outputs: Draft task for staff
Permissions: Draft only — no contact

Documentation Agent

Planned

Drafts notes for physician review.

Inputs: Encounter context
Outputs: Draft note
Permissions: Draft only

Data Quality Agent

Available in Demo

Validates inputs and flags issues.

Inputs: Raw datasets
Outputs: Quality report
Permissions: Read

Security & Governance Agent

Prototype

Checks access, PHI policy, and audit.

Inputs: Plan, user role
Outputs: Allow / block + audit entry
Permissions: Policy enforcement