Global consumer goods major Procter & Gamble (P&G) operates across more than 180 countries and manages a massive global supply chain, retail network, and portfolio of household brands. For a company operating at that scale, AI is increasingly becoming part of how the business runs.

To operationalize AI globally, P&G built a multi-layered AI organization that combines data science, AI engineering, and what the company calls its “AI Factory” model, a platform-focused capability designed to standardize, automate, and scale AI deployment across the enterprise.

According to Jeff Goldman, P&G’s Vice President of Enterprise Data Science, scaling AI requires organizational alignment, engineering discipline, standardized infrastructure, and deep integration with business operations.

“I came to P&G because I was excited about the possibility of applying AI at scale and operating on a global basis, applying algorithms to change how the company operates,” Goldman says.

In conversation with Donna Medeiros of Data Society Group, Goldman explains how P&G structured its AI organization, why the company created the AI Factory model, and what enterprises often misunderstand about scaling AI successfully.

AI’s long-term promise is finally becoming operational reality

For Goldman, the current AI wave is the continuation of a much longer technological trajectory. What has changed in recent years, he says, is not the underlying potential of AI, but the maturity of the technology required to operationalize it.

As computing power, data availability, engineering practices, and enterprise infrastructure improved, the scale of solvable business problems expanded dramatically.

“The business problems we can solve today are more interesting and complex than the ones we could solve a generation ago. We’re on a path to transforming how companies operate,” notes Goldman.

Why Enterprise AI Requires Both Local Relevance and Global Scale

One of the central themes in Goldman’s approach is balancing global consistency with local business relevance. That principle shaped how P&G structured its enterprise AI organization.

Rather than creating a standalone technical function disconnected from the business, the company built an operating model that combines:

• Data science innovation
• AI engineering
• AI production infrastructure
• Deep business integration
• Regional market alignment

Goldman explains that each layer serves a distinct purpose.

• The data science organization focuses on innovation and algorithm development.
• The AI engineering team focuses on scaling and operational reliability.
• The AI Factory organization focuses on standardization, infrastructure, automation, and production operations.

Together, the structure is designed to move AI from experimentation into repeatable operational capability.

The organizational design behind P&G’s AI strategy

Goldman says P&G’s current data science organization was established roughly 11 years ago using what he describes as a hybrid model. The goal was to ensure AI teams stayed closely connected to business priorities while also maintaining enterprise-wide coordination.

The structure includes:

1. Regional AI leadership

P&G has regional data science leaders embedded across major global markets.

“Their job is to embed themselves into the local business team, understand the priorities, and customize any algorithm we’re building to make sure they’re as locally relevant as possible.”

This helps ensure AI systems account for differences in markets, operations, consumers, and retail environments.

2. Global domain leadership

The organization also includes global leaders across three major domains: marketing, retail, and supply chain.

“Their job is to partner with the marketing and supply chain leaders of the company, transforming, building, and partnering on algorithms to transform their business processes.”

The global and regional structures work together to balance standardization with flexibility.

AI engineering became essential for scaling AI reliably

Goldman says one of the biggest lessons learned early in the company’s AI journey was that building models is only part of the challenge. Operationalizing them at enterprise scale is often significantly harder.

That realization elevated the importance of AI engineering.

The AI engineering team focuses on:

• Production deployment
• Reliability
• Scalability
• Integration with digital products
• Operational performance

Goldman credits the engineering teams with making complex AI initiatives operationally viable. Without strong engineering foundations, many AI projects struggle to move beyond experimentation into business-critical workflows.

Why P&G created an AI Factory

As P&G’s AI footprint expanded, Goldman says the company encountered a common enterprise challenge: fragmentation.

Different teams were building and deploying models in different ways. The infrastructure was inconsistent, and production processes lacked standardization. The inefficiencies became increasingly costly as AI adoption grew.

“There was a lot of time lost in provisioning environments, in replicating best practices and other things.”

In response, P&G launched its AI Factory initiative around 2021.

The goal was to automate operational overhead and create a scalable AI operating environment.

Goldman describes the initiative using the internal slogan: “Live life, model more.”

The idea was straightforward:

• Reduce operational toil
• Standardize AI deployment
• Automate infrastructure processes
• Increase collaboration

Goldman believes one of the biggest mistakes organizations make is treating AI as a technology-first initiative. “When you’re dealing with people working at AI, they often wanna use the most complex model or the most advanced analytic technique.”

But he argues that sophistication alone does not create value: “It should always start with a business problem that’s gonna deliver material value to the enterprise.”

To maintain that alignment, P&G intentionally embeds AI teams close to business operations.

Goldman explains that physical and organizational distance create misalignment.

“Every mile of distance separation between the people trying to transform the business algorithmically and the ones working on the business leads to disconnect.”

That proximity helps AI teams stay focused on operational impact instead of technical experimentation alone.

Funding structure creates continuous business accountability

One of the more notable aspects of P&G’s model is how AI initiatives are funded.

Goldman explains that much of the organization’s work is directly funded by business teams themselves. That creates an ongoing accountability mechanism.

Business leaders decide where they want to invest based on expected outcomes and operational priorities.

Goldman says this forces continuous prioritization around business value: “It gives a continuous reinforcement measurement every time we start a new project. Is this where I wanna invest my next dollar to deliver the maximum business value?”

AI at scale requires more than models

Throughout the conversation, Goldman repeatedly returns to one core idea: enterprise AI transformation is not primarily a modeling problem. It is an operational integration problem.

Algorithms alone do not create enterprise value unless they are embedded into business processes, scaled reliably, and aligned to measurable business outcomes.

“The best algorithm in the world, not integrated through a digital process, doesn’t deliver any value.”

For enterprises attempting to move beyond AI experimentation, that distinction may ultimately determine which organizations successfully operationalize AI at scale and which remain stuck in perpetual pilot mode.

CDO Magazine appreciates Jeff Goldman for sharing his insights with our global community.