AI Agent Operational Lift for Selee Corporation in Hendersonville, North Carolina

Why AI matters at this scale

Selee Corporation, founded in 1978 and headquartered in Hendersonville, North Carolina, is a specialty manufacturer of advanced porous metal and ceramic components. With an estimated 201-500 employees and revenues around $75M, Selee operates in a niche, high-value segment of the fabricated metal product industry. Their components are critical in demanding applications—chemical filtration, gas diffusion, and fluidization—where material integrity directly impacts customer process safety and efficiency.

For a mid-sized manufacturer like Selee, AI is no longer a futuristic concept but a competitive necessity. The sector is characterized by high-mix, low-volume production, skilled labor shortages, and thin margins on commodity lines. AI offers a path to differentiate through quality, speed, and cost control. Unlike large automotive or electronics plants, Selee likely lacks a dedicated data science team, making pragmatic, high-ROI use cases essential. The goal is not to replace craftsmen but to augment their expertise with data-driven insights, reducing variability in processes like sintering and inspection that are still heavily reliant on tribal knowledge.

Three concrete AI opportunities with ROI framing

1. Automated Visual Inspection for Zero-Defect Shipping The highest-impact starting point is deploying computer vision on final inspection stations. Porous metal filters require consistent pore size distribution; manual inspection is slow and subjective. A camera-based system trained on thousands of labeled defect images can flag cracks, contamination, or density variations in milliseconds. For a $75M manufacturer, reducing scrap by 15% could save over $500K annually in material and rework costs, with a system paying for itself within a year.

2. Predictive Maintenance on Critical Sintering Furnaces Sintering furnaces are the heart of production and a single unplanned downtime event can cost $50K-$100K in lost output and expedited shipping. Retrofitting furnaces with IoT sensors and applying anomaly detection models allows maintenance teams to schedule interventions during planned downtimes. This shifts operations from reactive to condition-based, potentially increasing overall equipment effectiveness (OEE) by 8-12%.

3. Generative Design for Custom Engineering Selee frequently produces custom filters to exacting client specifications. Today, engineers manually iterate designs. A generative AI tool, trained on historical successful designs and FEA simulation results, can propose optimized pore structures that meet flow and strength requirements in hours instead of days. This accelerates quote turnaround, a key competitive differentiator, and frees senior engineers for higher-value innovation work.

Deployment risks specific to this size band

Selee's size presents unique challenges. First, data scarcity: unlike mass production, custom manufacturing generates fewer, but highly complex, data points. Models must be trained on small datasets, requiring techniques like transfer learning. Second, talent gap: attracting AI talent to a mid-sized manufacturer in Western North Carolina is difficult; partnering with local universities or managed service providers is more viable than building an in-house team. Third, legacy integration: many shop-floor machines may lack digital interfaces, necessitating retrofits that can disrupt ongoing production. A phased approach—starting with a single, contained pilot on a non-critical line—mitigates these risks and builds organizational confidence before scaling.