Colgate-Palmolive

Colgate-Palmolive operates a sprawling global manufacturing network producing high-volume consumer goods — toothpaste, soap, and household products — where production continuity is critical to meeting retailer demand. Across its facilities, reactive maintenance practices left the company vulnerable to unplanned equipment failures that halted production lines without warning. In consumer goods manufacturing, even brief downtime cascades quickly into missed shipments, spoilage of in-process materials, and emergency repair premiums. The cumulative effect of these failures represented a significant drain on operational efficiency and maintenance budgets, with no systematic way to anticipate failures before they occurred.

Colgate-Palmolive partnered with Augury to deploy a continuous machine health monitoring platform across critical production assets. Augury's system combines vibration and ultrasound sensors mounted directly on rotating equipment — motors, pumps, compressors, and conveyors — with cloud-based predictive ML models trained on millions of machine-hours of industrial data. The platform analyzes sensor streams in real time, flagging early-stage degradation patterns weeks before they would cause failure. Maintenance teams receive prioritized alerts through a dashboard, enabling condition-based work orders rather than time-based schedules. The deployment followed a structured rollout across monitored lines, integrating with existing maintenance workflows without requiring replacement of legacy equipment.

Colgate-Palmolive achieved a 20% reduction in unplanned downtime within the program, with value demonstrated in as little as 3 months from initial deployment — a notably fast time-to-value for an enterprise-scale industrial AI rollout. The shift from reactive to predictive maintenance reduced emergency repair costs and allowed maintenance teams to schedule interventions during planned windows rather than responding to failures mid-production. Key outcomes included:

• 20% reduction in unplanned downtime across monitored assets
• 3-month time to value from deployment to demonstrated impact
• Improved Overall Equipment Effectiveness (OEE) on monitored lines
• Meaningful reduction in maintenance costs tied to emergency repairs

• Start with high-criticality assets: Targeting motors, pumps, and compressors with the highest failure impact accelerates ROI and builds internal confidence before broader rollout.
• Fast time-to-value is achievable: A 3-month payback window demonstrates that predictive maintenance doesn't require months of model training — pre-trained industrial ML models can surface value quickly.
• Sensor data quality determines model accuracy: Proper sensor placement on rotating equipment is foundational; shortcuts here degrade detection reliability.
• Change management matters as much as technology: Maintenance teams need workflow integration — alerts must map to actionable work orders, not just dashboards nobody checks.
• Condition-based scheduling reduces both cost and disruption: Shifting from calendar-based to condition-based maintenance minimizes unnecessary part replacements alongside unplanned failures.