Human-Centered AI Development Strategies for CPG Leaders – with Shaje Ganny of Procter & Gamble

CPG leaders need AI systems that start with concrete operating constraints and end with measurable outcomes vetted for company, consumer, and community impact.

MIT’s The GenAI Divide: State of AI in Business 2025 found that 95% of enterprise generative AI pilots did not produce measurable financial impact, a result that highlights how often companies launch AI before defining a specific business problem or operating model

At the same time, external analyses suggest that the potential value is significant when AI is integrated systematically.McKinsey found that, when AI is deployed across the full value chain, a $10 billion food-and-beverage business can generate $810 million to $1.6 billion in value, showing that the upside is significant when adoption is tightly linked to strategy and scale.

Consumer‑facing considerations add an additional layer of complexity. Washington State University researchers found that product descriptions using the term “artificial intelligence” can reduce purchase likelihood, and Consumer Reports found that 75% of Americans are concerned AI could lead to bias or unfair treatment in consumer-facing contexts.

Taken together, the data points to a consistent pattern. CPG firms are failing to translate AI experimentation into measurable value, despite large modeled upside. This gap emerges when initiatives are not clearly tied to business objectives, operating models, or consumer trust considerations.

Shaje Ganny, Digital Transformation Director at Procter & Gamble, joined Emerj’s Matthew DeMello on the AI in Business Podcast to explain how large CPG enterprises can responsibly scale AI using human‑centered operational principles.

This article examines three enterprise‑level capabilities Shaje Ganny argues are required for CPG companies to scale AI responsibly and reliably:

Problem‑defined AI operating models: Establishing AI around explicit business constraints creates the structural clarity needed for repeatable value capture rather than isolated pilots.

Three‑stakeholder impact governance: Integrating company, consumer, and community considerations into AI decisions ensures that automation strengthens the enterprise system rather than introducing risks to brand, safety, or the workforce.

Executive‑level AI fluency and accountability design: Building leadership competence in AI’s limits and responsibilities enables policies that prevent misalignment, operational gaps, and trust‑eroding deployment errors.

Listen to the full episode below:

Episode: Human-Centered AI Development Strategies for CPG Leaders – with Shaje Ganny of Procter & Gamble

​

Guest: Shaje Ganny, Author, Guest Lecturer, TEDx Speaker, and Digital Transformation Director at Procter & Gamble

Expertise: Enterprise AI, Digital Transformation, Digital Commerce, Strategic Planning

Brief Recognition: Shaje Ganny has spent more than two decades at Procter & Gamble, holding leadership roles across supply chain, demand planning, eCommerce, and digital transformation before becoming Group Director, Digital Transformation Europe. He leads digital transformation, digital commerce, master data, and joint business planning, including large-scale data integration with major retail partners. Previously, he was Global Digital Transformation Director for Digital Commerce, where he led P&G’s global eContent transformation, enterprise digital strategy, and early adoption of generative AI across multiple brands. Beyond P&G, he founded Swiss AI Academy, chairs the Education Sub-Committee for the IEEE Global Artificial Intelligence Systems Well-being Initiative, and authored AI Won’t Bite (2025). He holds a MicroMasters in Logistics, Materials, and Supply Chain Management from the Massachusetts Institute of Technology.

Problem‑Defined AI Operating Models

Shaje Ganny underscores that large CPG enterprises fail to scale AI when initiatives begin as technology experiments rather than responses to real operational constraints. In his view, AI only becomes repeatable and defensible when it is anchored to a specific business tension — line downtime, forecast volatility, quality‑control variability — where value can be measured and validated.

This shifts AI from discretionary experimentation to an operational requirement — forcing leaders to prioritize initiatives that resolve measurable bottlenecks over those that simply demonstrate technical capability.

He stresses that many senior leaders still misunderstand AI’s role, which leads to unrealistic expectations and stalled deployments. As he explains:

“Give me your vice presidents and presidents, and I will bring them fundamentals. They need to understand what AI is and what it is not. If they don’t understand the basics, they will ask for things that aren’t possible and then say AI doesn’t work. That’s not an AI problem—that’s a leadership problem.”

— Shaje Ganny, Author, Guest Lecturer, TEDx Speaker, and Digital Transformation Director at Procter & Gamble

Shaje’s interview surfaces a constraint‑first framing that helps executives distinguish scalable AI from novelty work. Leaders can operationalize this by requiring teams to specify:

The operational tension the AI is meant to address

The measurable variable that will indicate improvement

The accountable decision‑maker who validates value

The enterprise risk of not addressing the constraint

This structure provides leaders with a repeatable mechanism to evaluate whether an AI initiative is grounded in real business value and capable of scaling beyond a pilot.

Three‑Stakeholder Impact Governance

“CPG is an emotion-led business,” Ganny starts his explanation about why AI deployment in consumer goods cannot be evaluated solely through operational efficiency.

He stresses that consumers buy based on trust, story, and emotional resonance — and that AI today cannot replicate the authenticity that underpins brand loyalty. This makes the consumer lens as important as the operational one when evaluating AI’s role in the enterprise.

Ganny also highlights the broader system in which CPG companies operate: plants embedded in local communities, workforces with long‑standing skill bases, and brands whose reputations are built over decades. He warns that AI decisions made in isolation can create second‑order effects leaders often fail to anticipate. As he explains later in the conversation, failing to consider the workforce and community dimension introduces risks that “show up in ways leaders don’t expect.”

To make this lens concrete, Ganny points to three domains where AI decisions reverberate beyond the pilot environment:

Company: Reliability, safety, quality, and the ability to validate AI‑supported decisions

Consumer: Brand trust, emotional resonance, and the risk of perceived inauthenticity

Community: Workforce stability, local economic impact, and the social license to automate responsibly

This framing reflects the reality that CPG brands operate inside interconnected systems where operational, reputational, and social risks compound.

Ganny’s comments point to a Three‑Stakeholder Impact Scan as a practical way for leaders to evaluate whether an AI initiative is ready to move beyond the pilot stage. Before advancing any deployment, teams should document:

Operational impact — how the deployment affects reliability, safety, or quality

Consumer impact — whether the change could influence trust, authenticity, or brand perception

Community impact — how workforce roles, local employment, or community stability may shift

This approach reflects the interconnected system Ganny describes and gives executives a defensible way to determine whether an AI deployment strengthens the enterprise as a whole — not just the metric in front of them.

Executive‑Level AI Fluency and Accountability Design

When Shaje talks about AI in CPG, he doesn’t begin with algorithms or infrastructure; he begins with leadership. He argues that the biggest barrier to responsible AI adoption is not technical maturity but executive misunderstanding of what AI can and cannot do. Without that grounding, leaders unintentionally create expectation mismatches, governance gaps, and accountability confusion inside systems where safety, quality, and brand trust are non‑negotiable.

Ganny is direct about the stakes. In environments where a single decision can affect product integrity or plant safety, leaders cannot assume that responsibility shifts simply because AI is involved. As he puts it:

“I don’t know any CPG leader who would say they’re not accountable for the safety of their plant. But when AI comes in, suddenly people think the accountability shifts. It doesn’t. If the AI makes a decision, you are still accountable.”

— Shaje Ganny, Author, Guest Lecturer, TEDx Speaker, and Digital Transformation Director at Procter & Gamble

This point becomes sharper when contrasted with how AI is often deployed today. Many organizations treat AI recommendations as if they carry their own authority — a subtle but dangerous shift. Ganny stresses that AI cannot be allowed to “float” inside the enterprise without a clear owner, especially when it influences decisions tied to safety, quality, or consumer trust.

To prevent this drift, leaders need clarity on two fronts:

Decision boundaries — where AI may inform, where it may recommend, and where it must not decide

Accountability lines — who remains responsible when AI is followed, overridden, or fails

Ganny’s perspective makes it clear that AI maturity is measured by whether leaders understand and uphold these boundaries.

A practical takeaway from Shaje’s discussion:

Executives should establish an Accountability Map for every AI‑supported workflow. This includes:

The human owner is responsible for outcomes in that workflow

The decisions AI may influence, and the ones it may not

The escalation path when AI outputs conflict with human judgment

The conditions under which AI recommendations must be paused or reviewed

This approach ensures that as AI becomes embedded in daily operations, the enterprise does not lose sight of who is ultimately responsible for decisions that affect safety, quality, and brand trust.

​