Abstract:

Small and mid-sized Class II medical device manufacturers face a persistent operating challenge: regulatory evidence, supplier records, quality documentation, and operational decisions are often dispersed across functions, systems, and regions. This fragmentation is especially visible in diabetes device operations, where firms must manage FDA 510(k) submissions, EU MDR technical documentation, supplier qualification, design evidence, labeling control, and post-market expectations with limited organizational capacity.

This paper examines how governed artificial intelligence can help convert fragmented operational records into decision-ready knowledge while preserving auditability and human accountability. Using New Horizon Biotech as a case study, the paper develops and evaluates a dual-loop AI-enabled framework. The Regulatory Readiness Loop supports evidence retrieval, predicate comparison, submission drafting, and General Safety and Performance Requirements mapping.

The Supply Adaptation Loop supports supplier early warning, alternate-source readiness, change impact assessment, and total landed cost modeling. A six-month pilot using a five-FTE AI Business Unit showed measurable improvement across regulatory throughput, supply resilience, documentation quality, and financial value. AI-assisted regulatory drafting reduced cycle time from 6.5 to 4.4 weeks. Evidence retrieval time decreased from 12.5 to 1.5 hours. Lead-time recovery improved from 8.5 to 5.2 weeks, and sole-source exposure declined from 42% to 34% of the bill of materials.

The pilot also generated $1.55 million in gross value against an $850,000 investment, producing a positive six-month return. The study contributes to knowledge management, regulated operations, and supply chain resilience research by showing how a small, regulated manufacturer can use governed AI as an operating model rather than as uncontrolled automation.