From SAPinsider Las Vegas 2025: How Owens Corning Used AI-Powered Predictive Maintenance to Move to Reliability as a Service Model

Jeff Witt of Owens Corning shared the story of how the company used advanced condition monitoring and predictive maintenance to reshape asset management, reduce downtime, and generate substantial cost savings at SAPinsider Las Vegas 2025. Jeff and team have enhanced equipment monitoring in manufacturing using SAP Intelligent Asset Management and Asset Performance Management solutions. As part of the overall solution, the company has leveraged standardized equipment Information Models, cutting-edge predictive maintenance (PdM) technologies, and regional remote monitoring and diagnostic centers (RMDC).

The foundation of this transformation lies in wireless sensors that can be mounted anywhere in a facility and seamlessly connect to a centralized gateway. These sensors collect vibration, temperature, and pressure data in real time, feeding it into a cloud-based platform integrated with SAP Plant Maintenance (SAP PM).

This setup eliminates the need for manual monitoring, allowing the company to detect anomalies before failures occur, reduce unplanned downtime by transitioning from time-based to condition-based maintenance, and Enhance efficiency by integrating IoT data with business systems.

The APM (Asset Performance Management) module within SAP provides an analytics layer that alerts operators to potential failures, eliminating guesswork and enabling data-driven decisions.

A key challenge in industrial maintenance is identifying bad actors—recurring equipment failures that cause frequent yet minor inefficiencies. Traditionally, these failures might not be severe enough to warrant immediate attention, but when aggregated across multiple plants, they result in significant operational losses.

By leveraging automated alerts and condition monitoring dashboards, Owens Corning can now detect these small yet persistent issues, preventing widespread inefficiencies. For example, a minor one-minute stoppage per hour across 14 plants can add up to major productivity losses. With AI-driven alerts, these inefficiencies become visible, allowing teams to address them proactively.

The SAP-integrated solution enables seamless visibility into asset health by:

• Syncing IoT data with SAP PM, ensuring a single source of truth for maintenance teams.
• Automating alerts based on sensor thresholds, triggering notifications and work orders only when intervention is truly needed.
• Combining work history with real-time sensor data, allowing teams to optimize maintenance schedules and reduce unnecessary inspections.

One of the most impactful results has been the elimination of unnecessary preventive maintenance (PM) tasks, transitioning to a fully condition-based approach. Instead of maintaining equipment at fixed intervals, teams can now service assets only when necessary, leading to millions of dollars in annual savings per plant.

One of the biggest hurdles in deploying predictive maintenance solutions is false positives, where minor fluctuations in sensor readings trigger unnecessary alerts. Owens Corning tackled this issue by normalizing data to remove fluctuations caused by startup and shutdown cycles, implementing AI-powered alert filtering that ensures that only critical alerts reach human operators, and customizing thresholds at the plant level while maintaining a consistent corporate-level predictive model. These optimizations have dramatically reduced alert fatigue, allowing teams to focus on truly high-priority issues rather than being overwhelmed by excessive notifications.

By implementing AI-powered predictive maintenance across multiple plants, Owens Corning has achieved:

• $2 million in annual savings per plant by preventing unnecessary repairs.
• Significant reductions in emergency maintenance costs, thanks to early failure detection.
• Optimized workforce allocation, reducing the need for on-site staff while maintaining high equipment reliability.

The long-term goal is to integrate all IoT sensor data—including vibration, PLC (programmable logic controller) data, and historical maintenance records—into a unified AI-driven analytics platform.

What this means for SAPinsiders

AI-Enabled Condition Based Maintenance (CBM) is becoming best practice. By leveraging SAP IAM, APM, and PM with AI and IoT, organizations can transition from traditional, reactive maintenance to real-time, predictive maintenance. The key is to connect asset data, automate alerts, and use AI-driven insights to optimize maintenance strategies—ultimately maximizing uptime and reducing costs. Asset-intensive companies like Owens Corning should reasonably expect such benefits as 30-50% reduction in unplanned downtime, 10-30% cost savings from optimized maintenance schedules, 5-10% increase in asset life cycle due to proactive intervention, and 20-40% improvement in workforce efficiency with automated work orders.

Be aware of challenges with achieving an AI- and IoT-enabled CBM model. IoT sensors generate massive data streams, which can overwhelm systems if not properly managed. Poor data quality (e.g., inaccurate sensor readings, inconsistent data formats) can undermine AI/ML models and lead to false alerts or missed failures. Data integration challenges can arise when connecting IoT sensor data with SAP PM master data, leading to discrepancies between real-time and historical asset data. So it’s advisable to establish robust data governance policies to standardize sensor data collection, implement data cleansing and validation routines before feeding into SAP systems, and use SAP Business Technology Platform (BTP) and AI-driven anomaly detection to filter and analyze sensor noise. Also, CBM requires seamless integration between Operational Technology (OT) systems (SCADA, historians, IoT gateways) and SAP IT systems (IAM, APM, PM). Legacy systems may not be compatible with modern SAP architectures, requiring costly custom middleware. What’s more, cybersecurity risks increase when connecting industrial IoT devices to enterprise IT systems. So companies should use SAP Edge Services & SAP Data Intelligence to integrate OT & IT systems securely, implement SAP Cybersecurity best practices for IoT device security, and establish OT-IT collaboration teams to bridge communication gaps.

To maximize end-to-end CBM benefit, integrate with critical adjacent systems and processes. To enable real-time sensor data flow, companies are advised to integrate with SAP IoT (IoT-enabled asset monitoring), SAP Edge Services (Real-time processing for remote assets),  OSISoft PI System (Popular non-SAP IoT historian for manufacturing & utilities), and/or GE Predix, Siemens Mindsphere (Other IoT platforms for real-time equipment data). To link inventory and labor operations with CBM, companies should consider SAP S/4HANA Maintenance Management (to link CBM insights with work order execution), SAP Materials Management (MM) (for inventory of spare parts), SAP Controlling (CO) and SAP Financials (FI) (for maintenance cost tracking & budgeting), and SAP Human Capital Management (HCM) and SAP Field Service Management (for labor resource planning). To ensure that maintenance activities align with production schedules to minimize disruption, companies should integrate CBM systems with SAP Digital Manufacturing Cloud (DMC) (for intelligent production and asset tracking), Siemens Opcenter, Rockwell Automation MES, Honeywell PIMS (for real-time MES insights), and/or SCADA (Supervisory Control & Data Acquisition) Systems (for monitoring industrial control systems).