Few corporations can genuinely assert a century-long commitment to analytical research, yet Procter & Gamble stands as a prominent exception. Dating back to 1924, when CEO William Cooper Procter tasked economist Paul "Doc" Smelser with a groundbreaking study on Ivory soap usage, the consumer goods titan has consistently championed data-driven insights. Smelser’s findings – revealing that 40% of customers used Ivory for bathing versus 12% for dishwashing – fundamentally repositioned the product from a household staple to a personal care essential, a strategic pivot that underscores P&G’s early understanding of market segmentation and consumer behavior. This foundational legacy of rigorous analysis, coupled with pioneering efforts in establishing common data standards across a vast global organization, has laid the groundwork for the company’s ambitious foray into the age of artificial intelligence. Today, P&G leverages analytical, generative, and agentic AI to address critical business challenges, from product innovation to supply chain optimization, demonstrating a profound evolution of its data-centric ethos.
The company’s deep-rooted orientation towards data and analytics has not merely persisted but profoundly transformed with the advent of AI. Jeff Goldman, P&G’s Vice President of Enterprise Data Science and leader of business AI initiatives, emphasizes that this historical analytical drive now allows for an unprecedented depth of inquiry and a dramatically expanded ability to optimize business processes. Goldman, who previously spearheaded P&G’s innovative analytics visualization initiative, Business Sphere, now oversees a formidable team of several hundred data scientists and AI engineers. This specialized workforce is strategically embedded within business units and AI product teams, ensuring that AI algorithms are not just developed but deployed at scale across P&G’s diverse marketing, digital commerce, supply chain, and sales organizations, directly impacting operational efficiency and market responsiveness.
A pivotal development in P&G’s AI journey is the establishment of its "AI factory," conceived around 2021 in response to the increasing strategic importance and complexity of AI algorithms. Recognizing that bespoke development and custom deployment led to significant delays and financial impacts, the AI factory was designed as a comprehensive vehicle to accelerate the entire AI lifecycle. It provides a standardized ecosystem, encompassing data sources, robust software tools, validated methodologies, and stringent security protocols, enabling the rapid development, rigorous testing, seamless deployment, and continuous monitoring of algorithms in production environments. As P&G CIO Seth Cohen articulated, this platform approach grants developers instant access to both data repositories and advanced AI algorithms, including generative models, thereby eliminating the need to "worry about how to scale it" as that capability is inherent to the system. This strategic infrastructure investment has slashed model deployment times by an estimated six months, translating directly into faster market responsiveness and quicker realization of AI’s financial benefits.
The AI factory is not static; it dynamically evolves with technological advancements. Its current iteration, for instance, seamlessly integrates agentic AI capabilities, including the monitoring of agentic systems at scale, agent registration, and the implementation of sophisticated protocols like Agent2Agent and Model Context Protocol to facilitate interactions between multiple agents and tools. This dual focus, with one part of the AI Engineering organization building and continuously updating the factory, and the other scaling and operating the algorithms within it, ensures agility and efficiency. This framework facilitates rapid iteration and testing of different model versions tailored to specific business requirements. For example, the Pampers My Perfect Fit application, developed within this factory, leverages an AI-driven questionnaire to provide parents with diaper fit recommendations that boast 90% accuracy in preventing leaks – a critical factor influencing customer satisfaction and brand loyalty in the competitive baby care market. In another impactful analytical AI use case in Brazil, a system optimizes customer order fulfillment by intelligently splitting and scheduling truck-size loads based on shelf need, a solution that has remarkably reduced out-of-stock occurrences in the country by 15%, enhancing availability and sales.
While analytical AI continues to drive significant value in areas like supply chain management and media decisioning, P&G has also been an early and aggressive adopter of generative AI. The company has developed a suite of internal products, starting with chatPG, which provides employees with secure access to various underlying large language models tailored to specific business issues. This foundation has led to imagePG, a tool for generating and analyzing images and videos for advertising and other creative needs, and askPG, which integrates curated internal unstructured data to serve as an intelligent knowledge base for employees. These generative AI applications are not merely about personal productivity; they are designed to scale directly into core business processes. The Great Idea Generator tool, for instance, rapidly creates product and advertising concepts by analyzing consumer trends and past testing results, dramatically accelerating the journey from ideation to market. Similarly, Project Genie synthesizes information from extensive articles and help documents, empowering over 800 customer service representatives to reduce question processing time significantly, thereby improving customer experience and operational efficiency.
The evolution of data interaction within P&G is further exemplified by insightsPG, a generative front-end to business data. Building upon earlier initiatives like Business Sphere and Decision Cockpits, insightsPG reimagines how employees access and interpret data. As CIO Seth Cohen aptly questioned, "Why should you need a dashboard when you can talk to your data?" This innovation is transforming how the business interacts with its vast data repositories, democratizing advanced analytical and reasoning capabilities across a broader spectrum of the organization. P&G’s exploration of agentic AI is also progressing through a series of pilots aimed at identifying optimal deployment areas. Early successes have been observed across advertising, supply chain, and consumer relations, with Goldman emphasizing the critical importance of maintaining a "human in the loop" to ensure oversight and ethical deployment.
The synergy between Goldman’s data science and AI organization and P&G’s R&D function is particularly noteworthy. P&G’s R&D has a long history of employing quantitative analysis in understanding product chemistry, physics, and manufacturing processes. Now, AI algorithms are complementing the work of lab technicians, accelerating molecular discovery and product formulation. A prime example is the Perfume Development Digital Suite, an AI-powered ecosystem of digital tools that has cut the time to create new fragrances by five-fold. These algorithms analyze millions of data points and create intricate perfume character models based on consumer insights, identifying winning formulations that are then fast-tracked for prototyping and experimentation. This integration of AI not only speeds up innovation but also refines the development process, aligning new products more closely with consumer preferences.
Beyond technological implementation, P&G is actively reshaping its organizational culture and capabilities to fully embrace AI. A large-scale field experiment conducted in partnership with Harvard Business School’s Digital Data Design Institute explored AI’s role as a "cybernetic teammate." The study involved 776 commercial and R&D professionals tackling real consumer issues, revealing profound impacts. Individuals using chatPG achieved the same performance levels as teams without AI, while teams partnering with chatPG consistently produced superior outcomes. Crucially, the experiment found that AI helped dismantle functional silos, enabling professionals from both R&D and commercial backgrounds to produce more balanced and holistic solutions.
This commitment to human-AI collaboration extends to extensive upskilling initiatives across the workforce. P&G has partnered with Harvard Business School and Boston Consulting Group to deliver an intensive eight-week AI upskilling program, educating over 4,000 executives on the strategic impact of AI on the industry. Complementary in-house programs focus on integrating generative AI into daily work routines, fostering a rapidly growing cohort of leaders adept at piloting new AI capabilities and collaborating on the next generation of algorithms. Furthermore, the multi-year "Friends of Data Science" certification program continues to upskill P&G’s extensive quantitative analyst community, focusing not only on model building but also on understanding potential failure modes. The curriculum is continuously updated, now incorporating transformer models behind generative AI and graph machine learning for uncovering complex data signals. As CIO Cohen aptly summarizes, "A digitally reluctant organization makes it hard to introduce new capabilities. We’re changing that." Given P&G’s profound and sustained engagement with data, analytics, and now advanced AI, it stands as a global benchmark for how a legacy enterprise can not only adapt but thrive in the rapidly evolving digital economy. The strategic integration of AI across its value chain promises not just incremental gains but a fundamental redefinition of how consumer goods are conceptualized, developed, marketed, and delivered worldwide.
