Drawing on real-world applications from leading companies, the article demystifies how generative AI can be integrated not only for operational efficiency but for genuine customer-centric innovation.

Unlocking Customer Value through AI — An Analysis of INSEAD’s Six-Step Framework

by ChatGPT-4o

The INSEAD Knowledge article "Six Steps to Create Customer Value With AI" offers a compelling, pragmatic framework for how companies can move from AI experimentation to enterprise-wide transformation. Drawing on real-world applications from leading companies such as Diageo, Unilever, and Starbucks, the article demystifies how generative AI can be integrated not only for operational efficiency but for genuine customer-centric innovation. This essay unpacks each of the six steps, evaluates the strategic implications, and adds both critical reflections and additional recommendations.

1. Start Small (and Smart), Then Scale

The first step advocates a “pilot and scale” approach—launching low-risk, high-impact use cases to build momentum. Diageo’s partnership with Vivanda to develop the Flavor Print recommendation engine is a strong case in point. By aligning AI experimentation with unmet customer needs (navigating the complex whisky selection process), Diageo demonstrates how technology can enrich user experience in a relatable, emotionally resonant way.

Critique: While starting small is prudent, organizations must ensure that pilot projects are embedded within a larger transformation strategy. Isolated successes often fizzle out if not accompanied by strategic intent and long-term planning.

Additional Recommendation: Ensure early pilots include an exit strategy or success criteria tied directly to customer KPIs—not just internal efficiency or novelty.

2. Business Value First, AI Second

This step underscores the necessity of tying AI initiatives to clearly defined business outcomes. Diageo’s HALO project—allowing customers to co-create personalized whisky labels—demonstrates how generative AI can drive both customer engagement and revenue, yielding a 110% uplift in sales for Johnnie Walker Blue Label.

Critique: The emphasis on ROI-first is timely, especially in an era of AI hype where many firms are seduced by the technology itself. However, there’s a risk of overlooking longer-term brand equity or ethical consequences in the pursuit of immediate gains.

Additional Recommendation: Include ethical risk assessments in the value validation process—e.g., if a project drives revenue but deepens consumer manipulation or undermines trust, is it worth scaling?

3. Data as a Strategic Asset

No AI initiative thrives without high-quality data. Starbucks' Deep Brew platform is exemplary in operationalizing unified customer data to personalize experiences at scale. The case also highlights the importance of data lakes and cross-functional integration.

Critique: Many organizations underestimate the time, resources, and governance complexity needed to clean, standardize, and federate their data ecosystems. Cultural resistance to data sharing—especially in legacy environments—can sabotage AI ambitions.

Additional Recommendation: Invest in data literacy at all levels, not just among data teams. If your front-line workers don’t trust or understand data usage, the value chain breaks down.

4. Cross-Functional Collaboration

Diageo’s AI Council, which spans IT, legal, procurement, digital, and innovation teams, exemplifies how to institutionalize AI knowledge-sharing and risk mitigation. This collaborative governance model is essential to avoid “shadow AI” or siloed efforts that expose organizations to legal and reputational risks.

Critique: While councils are helpful, they often become bureaucratic or reactive unless they’re given executive-level sponsorship and resources.

Additional Recommendation: Formalize cross-functional AI charters—agreements on shared goals, principles, and responsibilities across departments—to reduce friction and accelerate decision-making.

5. Speed with Discipline: Agile Experimentation at Scale

This step emphasizes iterative, evidence-based innovation cycles. Diageo’s rollout of Seedlip, a digital brand ambassador, and its learning transfer to other brands is a strong use case for agile, modular innovation.

Critique: Agile methodology can falter when over-engineered or when executives demand “big bang” outcomes. Additionally, cultural inertia often undermines speed—even when frameworks are in place.

Additional Recommendation: Integrate AI experimentation into broader enterprise OKRs (Objectives and Key Results) to institutionalize a culture of continuous testing and learning.

6. Technology Stack Scalability

A flexible, modular (headless) tech stack enables companies to remain vendor-agnostic and adapt to evolving AI models. Unilever’s success in decoupling front-end tools from back-end systems allows seamless integration of best-in-class solutions across geographies and functions.

Critique: Headless architectures can introduce integration challenges, technical debt, and cybersecurity risks if not managed carefully. Many organizations underestimate the architectural discipline needed to avoid Frankenstein-like stacks.

Additional Recommendation: Mandate a “composability assessment” before introducing any new tool—will this new tool integrate cleanly into the stack, or will it create brittle dependencies?

Guiding Principles: The “How” Behind the “What”

The six execution steps are underpinned by essential principles:

• Evidence-based decision-making

• Rapid experimentation

• Top-down sponsorship with bottom-up empowerment

• Customer-centric design

• Trust and ethical alignment

These are not just platitudes—they form the cultural and strategic bedrock for successful AI adoption. However, the final principle—“Build for trust”—is arguably underdeveloped in the original article.

Expanded Perspective on Trust: With generative AI capable of hyper-personalization and deepfakes alike, businesses must go beyond transparency to earn durable trust. That includes:

• Explaining AI decisions in human terms

• Offering opt-out mechanisms for data-driven personalization

• Aligning AI usage with societal norms and regulatory frameworks (e.g., GDPR, EU AI Act)

Additional Strategic Recommendations

1. Introduce “AI Shadow Boards” – Engage younger or non-technical employees to provide input on how AI tools will be perceived by less digitally-savvy customer segments.

2. Adopt an “AI Rights Charter” – Codify principles that protect customers from misuse of AI, including misuse of data, deceptive personalization, or exclusionary algorithmic bias.

3. Benchmark Against Ecosystem Peers – Continually assess your AI roadmap against competitors, partners, and market disruptors. AI leadership is highly perishable.

4. Decentralize Experimentation, Centralize Governance – Encourage local teams to experiment but ensure unified frameworks for risk, compliance, and shared learning.

Conclusion: AI as a Value Engine, Not a Vanity Project

The INSEAD framework succeeds by balancing ambition with discipline. Its six steps chart a path for companies to move beyond the pilot trap and toward scalable, customer-centric innovation. Yet to succeed in today’s complex AI landscape, organizations must add greater attention to ethical foresight, trust-building, and cultural transformation. AI is not just a toolset—it’s a mindset. Those who embed it across their strategy, systems, and soul will thrive.

Now is the time to act—not merely to adopt AI, but to adapt to the world it is rapidly reshaping.