Neurodevelopmental care is often slowed by fragmented intake, long queues, and uneven coordination across clinics, schools, and families, which extends time-to-referral (TTR) and deepens inequities. We present AURA-T (Autism Universal Rapid Assessment Tool), a standards-first, privacy-preserving pipeline that moves from external form intake (Tally) into an operational FHIR store (Azure Health Data Services—FHIR R4/5), then to a de-identified analytics plane (FHIR Bulk Data $export to ADLS Gen2 with views in Synapse), and finally to retrieval-augmented report generation and secure delivery (Azure AI Search + Azure OpenAI; single-use links via Azure Communication Services). The design enforces least-privilege access, end-to-end Provenance/AuditEvent, and a strict split between PHI-bearing operations and downstream analytics/LLM stages, with US Core/USCDI alignment and compatibility with Brazil’s RNDS (Rede Nacional de Dados em Saúde). Current U.S. surveillance estimates about 1 in 31 eight-year-old children with autism in 2022 network communities [1], while global estimates are near ~1 in 100 children [2]. Because this venue emphasizes complex-network methods, we frame FHIR-as-graph analyses as methods-only (no outcomes claimed): multiplex modeling of coordination (clinical, school, family/caregiver, logistics layers), structural controllability, robustness/percolation, and referral link prediction with calibrated decision policies. A 90-day PoC targets payer/provider scale, ships ANS-first PDFs with embedded provenance, and defines KPI templates to evaluate operational value under governance. (Standards: US Core [3]; RNDS [4, 5]).