Team 58

Thalamus

Thalamus is a nursing personal agent platform that helps behavioral-health hospitals connect patient and payer context, prior authorization, continued-stay review, and denial recovery in one human-reviewed case record.

Check your own project

Community votes recognize the build and do not affect official judging.

Review the project

Start with the source code, then open the demo or video if available.

View GitHub repository
Project description
Thalamus agent platform gives every behavioral-health nursing and utilization-review professional a dedicated AI teammate, spanning the whole insurance journey from eligibility check to recovery. It unifies patient and payer context, prior authorization, concurrent review, and denial recovery in a single case record. Staff no longer have to chase status across portals, faxes, and phone calls. The process is AI-assisted but expert-reviewed, and it is designed for behavioral-health complexity. Architecture comes with, Unified Case Schema: EHR data (FHIR/HL7), payer portals, clearinghouse transactions (X12/EDI), documents, and call/fax inputs are normalized into one structured case. Each case has a single owner and a continuous state. Specialized agent layer: Orchestration, Document, Integrations, Medical-Necessity, UM/PA, Denial, and Appeal agents, plus an AI Evaluation agent, collect evidence and draft payer-ready packets. Behavioral-health rules engine: It encodes ASAM, LOCUS, MCG, and 42 CFR Part 2 considerations, so the output follows payer criteria and consent rules. Act-or-Defer safety: The system should not guess when evidence is weak. Low-confidence cases are routed to human review instead of being submitted. Auditability and security: Every AI judgment, source, and human approval is logged. The platform runs on AWS and has a SOC 2 Type 1 audit report. Thalamus works inside each user's own cases and keeps them in individual context.
Tools used