Proctor & Gamble (P&G) maintains an enormous product assortment, which it classifies into ten product categories, like oral care, hair care, and skin and personal care. Despite some links and commonalities across these categories, each of them raises different demands and requirements, so when it comes to performing marketing analyses, the corporation needs different sources and types of information. In particular, it increasingly relies on AI-enabled sources of information, and in this effort, it displays its readiness to interact with diverse partners that can support its diverse efforts across its diverse product categories and product lines.

At a company-wide level for example, it partnered with OpenAI. Initially, it used the company’s existing AI tools, but over time, recognizing the potential threat of intellectual property spillovers, P&G worked with OpenAI to build its own, proprietary system, called chatPG. The system mimics many of the capabilities of the widely available ChatGPT, but it is completely internal, such that none of the data that P&G feeds into chatPG are accessible to outsiders. With this system, P&G encourages employees to participate in what it calls its “AI factory,” by uploading challenges, questions, and use cases. The responses and suggestions produced by chatPG in turn can reveal novel solutions and information about previous trials and errors, leading to more efficient, effective, strategic decision-making. 

For more specific uses though, P&G turns to subject experts. Many of its products—including detergents, shampoos, deodorants, and cleansers—base much of their appeal to consumers on the beautiful or comforting scents they provide. But how can P&G design new scents for its products? An existing AI system called Moodify White promises assistance. The company owns a massive fragrance database, to which the AI applies various algorithms to come up with appealing scent combinations, quickly and with little additional cost. Through this collaboration, P&G expects to introduce new combinations of scents to the market more quickly, confident that the proposed scent profile has been established as attractive to consumers.  

This approach also has been informed by yet another collaboration, with the consulting firm Accenture. By applying its own AI, Accenture developed an extensive plan for how to embrace a “formulation” approach to new product designs. That is, P&G products comprise vast numbers of features, including scent but also cleaning power, color, and so forth. A traditional formulation process would involve hundreds of human employees considering and testing the virtually innumerable combinations of features. By detailing a company-specific “Human + Machine” framework for how to integrate AI into this process, Accenture offered P&G a plan it could apply not only to new product development but also across its functions, such as operations and human resources.

Before consumers can enjoy these features though, they have to be convinced that they need the products that contain them. In a marketing effort, P&G teamed with manufacturers of smart appliances to gather data facilitated by the Internet of Things (IoT). For example, with data from smart dishwashers, the company determined that today’s machines offer greater water and energy efficiency than washing dishes by hand. In turn, P&G leveraged this particular form of AI to establish an environmentally oriented campaign designed to convince consumers to run their dishwashers, even if they were not completely full, rather than fill a sink with water. 

The company’s other environmental efforts similarly are facilitated by AI systems installed in its factories. As detailed by the P&G’s chief information officer, “We extract data from our production lines with sensors, sending it to the cloud to build and train machine learning algorithms that are then deployed back to the edge on the factory floor. Thanks to this approach, we can decrease overpacking of paper towel rolls, for example, and automate quality controls for feminine pads in real time using computer vision.” Even if P&G combines all these technologies together within its factories, it looks outward, to partnerships with various external technology providers, to obtain them. 

Discussion Questions

• Should P&G’s competitors collaborate with OpenAI on their own systems? Use an existing system? Partner with another firm? 
• How might AI inform product formulations in other market sectors? 

Sources: Lindsey Wilkinson, “How P&G Rolled Out Its Internal Generative AI Model,” CIO Dive, October 18, 2023; Katherine Noyes, “Proctor & Gamble CIO: ‘Everybody Must Own the Digital Story’,” CIO Journal, January 27, 2023; Accenture, “Brand New: AI-Tailored Product Development,” https://www.accenture.com/us-en/case-studies/consumer-goods-services/pg-ai-tailored-product-development; “Proctor & Gamble Partners with Moodify to Create AI-Based Fragrances,” Premium Beauty News, October 16, 2023; Proctor & Gamble, “A Strong and Focused Portfolio,” https://us.pg.com/annualreport2021/a-focused-portfolio/