Augmenting the Capabilities of the Enterprise by Scaling AI

Picture this: a P&G sales representative visits a retail store to review the performance of P&G products. Instead of having to capture data on her own, such as the accuracy of shelf layout, completeness of the assortment or execution of promotions, the representative can access this information directly on her mobile device, along with alerts recommending specific actions to improve performance and topics for discussion with store management.

Those alerts and recommendations are produced with granular accuracy by an AI engine working on large amount of data collected from diverse sources. This is just one example of how using data and AI to produce predictive insights and prescriptive recommendations can significantly augment P&G’s business performance and deliver superior experiences to billions of consumers around the world.

What has made AI a true differentiator is the ability to go beyond one-off initiatives to systematically scale algorithmic solutions in multiple company categories and markets around the world. When scaled pervasively across the organization. AI can reach its true potential to exponentially augment the capabilities of the enterprise.

It is my belief that to make scalability a reality, businesses must invest in three key areas: a robust data platform, strong talent and skills, and automation.

Like many large companies, P&G had a legacy data architecture, with large data siloes that, while useful to specific functions, did not enable us to easily gather, share and reapply learnings across the business. To break these silos and enable use cases across the company, we built a cloud-enabled data lake to consolidate data, complement them with metadata and aligned taxonomies, and made them available through user-friendly catalogs. This was a necessary step because there aren’t yet data mesh solutions that can be applied at scale, connecting data regardless of where they sit.

For these data points and insights to be meaningfully applied to achieve the highest impact, it’s critical to have smart, talented people who understand how to analyze and work with the information. At P&G we have hundreds of data scientists who are organized as a team to foster sharing of best practices and optimal career development. They are embedded across the organization to ensure they have proximity to use cases. ?We also invest in strong internal data and machine learning engineering specialists, with a healthy 50/50 ratio of internal vs. external resources to maintain expertise and the ability to effectively manage third party vendors.

To ensure our teams are developing their skills, we are consistently investing in the AI fluency of the organization. For example, we offer an internal training program with two levels of certification, including a 15-week class on fundamentals of data science. ?And, thanks to our partnership with the Harvard Business School, several hundred of our business leaders have attended a course on “Competing in the Age of AI,” which focuses on rethinking how businesses compete and operate in the era of AI.

The final key to AI scalability is an investment in the latest tools for building AI products. These tools minimize the toil involved with the creation of data pipelines. Our investment includes an “AI Factory” workbench which simplifies the end-to-end process of creating, training deploying and monitoring an algorithm. This increased level of automation allows for a broader group of players to develop AI solutions, while improving model risk management and observability.

The world is constantly changing. Our customers face old and new challenges every day, and we keep innovating to provide them with superior solutions for those. Technology is an essential enabler of that mission, and AI is a pervasive capability across P&G. As we continue our journey, we are committed to investing in automation, maximizing the impact of our talented people and accelerating time to value of algorithmic solutions.??