Medline Industries

Andrew Splawn is a skilled Multi-site Global Plant Manager and Certified Manufacturing Specialist with extensive experience in the medical device industry. Currently managing operations for Medline Industries in Tyler, Texas, he is known for driving efficiency by building effective leadership teams, applying lean manufacturing techniques and Six Sigma practices.

Leveraging his deep experience in operational excellence, Splawn recognizes that the next frontier in manufacturing efficiency lies in digital transformation. Through this how-to article, Splawn provides a two-phased approach for leveraging AI tools to transition from outdated manual scheduling (using paper and Excel) to an AI-driven, predictive operational model.

The transition from outdated manual scheduling to an AI-driven, predictive operational model is a critical step for modern manufacturers aiming to reduce costs and maximize throughput. This journey can be achieved through a strategic, two-phase approach, realizing quick wins before making capital investments for long-term optimization.

Phase 1: Improving Your Preventive Maintenance Program (Free/Low-Cost AI Automation)

The first phase focuses on leveraging accessible Generative AI to instantly clean, standardize and optimize your current time-based Preventative Maintenance (PM) and production schedules. This step improves the quality and accessibility of your current content.

To begin improving your current paper-based processes using readily available AI tools, follow these steps:

• PM Schedule Generation: Leverage AI to quickly generate standardized, comprehensive PM checklists and schedules based on equipment specifications. Prompt the AI with a request such as: “Create a 52-week time-based PM schedule for a centrifugal pump, including weekly checks, monthly lubrication and quarterly sensor calibration. Output the result in a clean, spreadsheet-friendly format.”

• Task Prioritization: Use AI to analyze any backlog work orders and current production jobs (copied from your Excel sheet) to suggest the optimal execution sequence. Prompt the AI with a scenario like: “Given this list of 20 maintenance work orders with required duration, necessary skill level and a fixed production due date, generate a prioritized daily task list for the maintenance team to minimize overtime and late deliveries.”

• SOP/Documentation: Employ AI to convert existing handwritten notes or poorly formatted guides into clear, digital Standard Operating Procedures (SOPs). Prompt the AI with an instruction such as: “Turn this text [insert current process] into a five-step, easy-to-read checklist for performing a die change on the press, highlighting all safety requirements.”

Once the AI generates these schedules and SOPs, they must immediately replace the paper copies. Transfer the finalized documents to a shared digital platform (e.g., Google Sheets, OneDrive, or Microsoft Planner).

“By focusing on critical assets and proving the ROI of preventing those costly failures first, your organization can successfully build momentum for a facility-wide predictive strategy.”

Bonus Step (low cost): introduce a system of digital feedback. Use simple QR codes placed on equipment that link directly to the new, AI-generated PM checklists. Maintenance scans the code, executes the task, and digitally marks completion. This simple step replaces clipboards and creates a foundation of clean, time-stamped digital completion data, which helps transition into Phase 2.

Phase 2: Predictive Maintenance (Capital Investment & Integration)

This phase requires strategic capital investment to gather real-time data, enabling a shift from scheduled PMs to Condition-Based Monitoring (CBM) and dynamic scheduling, which drastically reduces unexpected downtime.

1. Strategic Asset Selection: Maximize ROI

Identify the equipment that provides the most value and whose failure causes the highest cost of downtime. This is your starting point for sensor investment.

• Prioritize Consequence: A massive, mainline compressor or a central furnace might fail rarely, but its failure is catastrophic, stopping the entire production line. Start with the catastrophic failure point.

• Identify Failure Modes: Determine how the equipment typically fails (e.g., motor bearing wear, fluid contamination). This dictates the necessary IoT sensor (e.g., vibration analysis for bearing wear, temperature sensor for overheating).

2. The Data Pipeline and Infrastructure

The key to predictive maintenance is the data flow:

• IoT Sensors: Wireless sensors are installed to gather high-frequency data (vibration, temperature, current, acoustics).

• Edge Computing: A small computer at the facility processes the massive stream of raw data locally, converting it into manageable features, reducing network load.

• Data Historian/Lake: The processed data is transmitted and stored alongside historical data (past failure reports, operating hours) in a dedicated Data Historian or Data Lake. This structured repository is the necessary fuel for the AI model.

3. AI For Prediction And Dynamic Scheduling

Once the data is clean and integrated, the AI engine can be applied:

• Predictive Maintenance (PdM): Machine Learning models analyze the real-time sensor data to detect subtle anomalies and calculate the Remaining Useful Life (RUL) of critical components.

• Dynamic Advanced Planning & Scheduling (APS): The APS system receives the PdM output (e.g., “Pump A bearing failure risk hits 85% in 8 days”). The APS automatically:

1. Generates a critical work order.

2. Re-runs the optimization algorithm to insert the maintenance task into the production schedule at the optimal time, ensuring the intervention causes the least possible disruption to high-priority customer orders.

Finally, adopting AI requires a cultural shift. Maintenance, Production and IT teams must collaborate. Technicians must be upskilled to trust and act upon AI-generated condition alerts rather than waiting for a catastrophic failure. By focusing on critical assets and proving the ROI of preventing those costly failures first, your organization can successfully build momentum for a facility-wide predictive strategy.