Case Study

HealthTech Startup Scales from Prototype AI Tool to Enterprise-Ready Clinical Platform Serving 250K+ Patients with AWS Bedrock and SageMaker

At a glance

Health insurers cannot adopt clinical AI from startups without a production-grade, auditable platform underneath it. For this Houston-based digital health startup, that ceiling was existential: without enterprise-ready infrastructure, the company could not close the government contracts it had already earned, including a Southeast Asian national deployment targeting 4 million patients.

NMD built the client’s clinical AI platform on Bedrock and SageMaker, automating guideline-based triage, risk scoring, and referral generation across 250,000 patients. Coordination cycles that took days now complete in minutes, a 95%+ reduction, with the auditable outputs enterprise health systems require.

The company is now onboarding its first enterprise clinic customers. AWS consumption is projected to scale from under $3K to $32K MRR post-deployment, 10x growth with $21K driven by SageMaker. A $2.5M to $3M AI EHR platform is in active scoping, backed by a $10M US Export-Import Bank Letter of Intent.

Industry

Use Case

Solution implemented

The value equation

Company Snapshot

A Houston-based HealthTech startup delivers AI-powered clinical decision support to care teams managing large patient populations, enabling health systems and government payers to apply evidence-based protocols, identify high-risk patients, and coordinate care at scale.

Location

Houston, TX, USA

Customer Situation

This Houston-based digital health startup holds patented technology developed in partnership with a leading US research university. Health systems broadly lack auditable tools to enforce care pathways and coordinate referrals across fragmented EMR environments. Manual workflows delay intervention, increase denied claims, and drive preventable emergency visits.

The barrier was enterprise readiness. Regulated markets require HIPAA-aligned, multi-tenant platforms validated at population scale. The company had a strong proof of concept but could not clear that bar, blocking the government contracts that represented its primary growth path.

A Southeast Asian national health insurer selected the platform to replace its sunsetted EMR infrastructure across a multi-million patient population, backed by a $10M US Export-Import Bank Letter of Intent. An internal build would have lost that government window entirely.

NMD Solution

NMD identified that the gap between the client’s proof of concept and enterprise adoption was an architecture and compliance problem, not a clinical AI problem. The intelligence logic existed. What was missing was the regulated-industry infrastructure layer that enterprise health systems require: multi-tenancy, PHI audit logging, guideline-based rule pipelines with clinical sign-off, and ML-validated risk scoring at population scale. Closing that gap required a production-grade AWS build: SageMaker for population-level risk stratification, Bedrock with Claude for clinical protocol reasoning and agentic care workflows, and HealthLake providing the FHIR-native data layer that enterprise EMR integrations demand.

What We Delivered

NMD built the clinical AI platform in 8.5 weeks through AWS Activate-funded delivery, compressing an 18-month internal timeline. Three modules were delivered: a Health Concierge AI for benefit-guided navigation and scheduling; a Referral Integration engine routing patients by condition and eligibility; and a Chronic Disease Management pipeline scoring risk via SageMaker and delivering prioritized care alerts. Validated across 250,000 patients, coordination cycles that took days now complete in minutes, a 95%+ reduction, with the audit trails enterprise and government clients require. AWS MRR is projected to scale from under $3K to $32K, with $21K driven by SageMaker.

Ready to Transform Your HealthTech Business?

See how New Math Data can transform your healthtech business with AWS-powered innovation.