How Domtar turned AI sensor data into a maintenance advantage

The breakthrough came when McLaughlin shifted the company’s relationship with vendor Waites Sensor Technologies from passive monitoring to active partnership. By moving from monthly to weekly collaborative calls, Domtar began feeding a broader spectrum of diagnostic data—including thermal imaging and ultrasonic metrics—into the machine-learning system. This feedback loop allowed the AI to move beyond generic alerts, refining its predictive accuracy for specific machinery health.

To bridge the gap between algorithmic output and plant-floor reality, McLaughlin implemented a daily reporting system for legacy staff. By demonstrating the system's reliability through consistent, trackable results, he fostered enough organizational trust that managers now rely on the data for major capital decisions. This cultural shift, paired with the increased sensor count from 450 to 748, helped the facility eliminate unplanned downtime by over 1,500 hours. The impact is visible in mundane metrics: the mill has not needed to purchase replacement fan belts for an entire year, a direct result of the system’s ability to detect and rectify minor misalignments before they lead to catastrophic failure.