Procter & Gamble (P&G), a global leader in consumer goods, has been at the forefront of integrating modern technology into its operations. With over 120 manufacturing sites in over 40 countries, the company’s digital journey commenced approximately seven years ago. Jeff Kent, the vice president of Smart Platforms Technology & Innovation, along with other key leaders, has been steering the company towards full-scale digitalization.
Procter & Gamble has acknowledged the profound potential of AI and machine learning as pivotal elements in production digitalization. This recognition paved the way for the inception of the WISE initiative and the deployment of the SmartBox edge device, central pillars in the company’s AI implementation strategy. Alongside these advancements, the company instituted the Manufacturing Machine Learning (ML) Lifecycle, a structured approach designed to seamlessly integrate and optimize AI applications in their operations.
Procter & Gamble employs the SmartBox device to amass data from a myriad of controls and equipment, ensuring that information flows in real-time and is seamlessly integrated with cloud applications. This adoption of technology is accentuated by the company’s collaboration with industry forerunners, such as Microsoft, in their endeavor to craft robust AI models. In a notable application within Pampers manufacturing, these AI models are adeptly trained using an amalgamation of both genuine and simulated data, maximizing their efficiency and accuracy.
Procter & Gamble has witnessed remarkable advancements in its manufacturing operations, most notably an impressive 80% reduction in glue-related waste during the production of Pampers. This transformation is bolstered by the incorporation of real-time alert systems, allowing for an immediate response to any manufacturing anomalies, thereby ensuring the consistent quality of the product.
In P&G’s journey towards manufacturing excellence, certain challenges have emerged. Although production errors had high associated costs, they occurred infrequently, resulting in their underrepresentation in the training data. This posed a significant hurdle for the AI-driven approach. Additionally, the limitations of the traditional Purdue model in the system architecture necessitated a move towards a more open and communicative structure. One of the intricate challenges included maintaining flawless consistency in intricate processes, such as the 40 distinct glue streams essential for manufacturing Pampers.
P&G is set on transforming its architectural approach by moving away from the constraints of the traditional Purdue model, aiming for a more open and interconnected system. Central to this transformation is their continued emphasis on OPC UA, recognized for its crucial role in bridging the communication between OT and IT. Ambitiously, P&G has its sights set on implementing these advancements across all 120 of its factories within the span of a year.