AI Lessons From Dutch Boardrooms: What Ahold, ING, KLM and NS Teach Us About Enterprise Transformation

by ChatGPT 5.1

The article “AI lessons from the boardrooms of Ahold, ING, KLM and NS” offers a rare, candid look into how some of the Netherlands’ most influential corporate leaders approach artificial intelligence. Their experiences reveal a set of strategic, organisational, and cultural lessons that any enterprise—especially those in traditional, regulated, or infrastructure-heavy sectors—can learn from. What emerges is not a story of hype, but one of focus, pragmatism, and disciplined experimentation.

Across retail, banking, aviation, and public transportation, the same themes recur: AI only works when data foundations are strong, when implemented in coherent processes rather than scattered pilots, and when humans are invited—not coerced—to adopt new tools. These insights carry wider implications for global enterprises navigating AI’s shift from experimentation to scalable impact.

1. From Hundreds of Pilots to a Coherent Strategy: The Ahold Delhaize Lesson

Ahold Delhaize’s journey mirrors that of many large organisations: initial euphoria, countless pilots, and eventually, the realisation that scattered innovation does not equal meaningful transformation. The company tested everything—from food-waste reduction to logistics optimisation—and even achieved measured benefits such as a 6% drop in kilometres driven through smarter route planning at Bol.

Yet CEO Frans Muller recognised a deeper problem: too much fragmentation, too little strategic focus.

Instead of “hundreds of small projects,” Ahold is now concentrating on two or three end-to-end processes, such as supplier negotiations or the full supply chain. The idea is to redesign entire workflows, not just apply isolated tools. Muller stresses that the “foundation is data and data quality”—a theme echoed by every other CEO in the article.

What makes Ahold’s stance especially pragmatic is its self-identification as a “smart follower”. Ahold knows it doesn’t have to invent everything; if a larger or more specialised company produces a solution, Ahold will adopt it instead of reinventing it.

This is a strategic humility often lacking in companies that either overestimate their technical capabilities or underestimate the value of buying.

2. Technology Is Easy—People Are Hard: The NS Lesson

NS CEO Wouter Koolmees is refreshingly honest about AI’s duality: part hype, part inevitable transformation. His caution derives from past experience: AI tools fail not because they don’t work, but because humans don’t adopt them.

NS learned this the hard way. A predictive maintenance system was technically sound, but engineers refused to follow it, trusting their own 20 years of experience. The issue was not the model—it was how it was positioned. It felt like “the system says no,” a threat to autonomy.

When the tool was reintroduced as an assistant, not a master, adoption skyrocketed.

The same principle underpins NS’s later success with AI-assisted anomaly detection using trackside camera poles, which let staff inspect train components during the day rather than crawling underneath trains at night. This saved labour time, reduced night shifts, and relieved pressure amid workforce shortages.

NS also demonstrates strategic selectivity. It doesn’t try to lead in generic AI tools (chatbots, Copilot) but aims to innovate in logistics and safety, where the railway’s complexity gives it a competitive advantage.

3. Start With the Problem, Not the Algorithm: The ING Lesson

ING has one of the longest digital transformation histories in Dutch industry, and its AI strategy reflects that maturity. Rather than letting innovation drift into “cool things without business impact,” CEO Steven van Rijswijk places AI ownership within operations, not IT, and assigns accountability to business leaders.

The process always begins with a simple question:
What problem are we solving?

This mindset prevents misalignment between technology and business objectives.

When ING tested a generative AI chatbot on 1,000 customers, it did so knowing that “it would occasionally say strange things”—in fact, the abnormal behaviour was part of the experiment. Only real-world deployment could reveal underlying data and process issues.

Another ING lesson: context matters. When the Dutch chatbot was deployed abroad, it performed poorly because regulations differ (e.g., Dutch law prohibits offering financial advice via chatbots; Spanish law does not). This required country-specific redesign.

One of ING’s strongest examples of AI value is in anti-money-laundering. AI now filters obvious false positives—90% of the bank’s millions of annual alerts—allowing human investigators to focus on real cases.

Their philosophy is clear: avoid dozens of small pilots; concentrate on a handful of high-impact processes.

4. From Legacy AI to Generative AI: The KLM Lesson

KLM has been using AI since the 1980s in cockpit systems, and since 2015 with tools like Pathfinder, which optimises aircraft positioning. But generative AI represents a new frontier.

CEO Marjan Rintel openly admits that even the boardroom needed to “learn what AI can do,” challenging the corporate myth that executives always understand the technologies their organisations deploy.

KLM’s strategy resembles Ahold’s and ING’s: move from scattered pilots to a focused, coherent AI roadmap. Their successes—meal-planning prediction to reduce food waste, fuel optimisation systems, and AI-assisted technical documentation—are examples of AI that enhances efficiency while respecting the human element.

Rintel does not believe aviation will be radically automated soon. With 80% of airline work still human-driven, particularly during disruptions, AI’s role is assistive rather than transformative.

Remarkably, KLM is even commercialising its AI planning tools for other airlines—a rare case of a traditional company productising its internal technology.

5. Collaboration as a Competitive Advantage: KickstartAI

Perhaps the most interesting structural insight from the article is the joint initiative KickstartAI, through which Ahold, ING, KLM, and NS pool talent and work on societal AI projects such as a Food Bank logistics tool.

This collaboration:

• accelerates innovation

• spreads best practices

• combats talent retention challenges

• connects private sector AI to societal goals

It shows that AI ecosystems outperform isolated enterprises—a lesson useful for governments and industries across Europe and beyond.

Conclusion: The Emerging Dutch AI Playbook

Across four different industries, a clear pattern emerges—a practical, grounded AI transformation philosophy:

1. Start with data quality and process architecture.

Without clean data and well-defined workflows, AI will fail.

2. Replace scattered pilots with a focused AI portfolio.

Pick 2–5 high-impact processes; stop the “pilot chaos.”

3. Position AI as a collaborator, not a dictator.

Human adoption determines success more than model quality.

4. Embed AI ownership in the business, not only IT.

Technology must align with operational realities, not technical fascination.

5. Prepare leadership—AI fluency must start at the top.

Executives must understand what AI can and cannot do.

6. Accept regulated constraints as design parameters.

Compliance is not a barrier but part of the system.

7. Collaborate across institutions.

Shared initiatives like KickstartAI exponentially increase talent and impact.

Final Thoughts

This article captures a European sensibility toward AI: pragmatic, human-centric, and grounded in process improvement rather than grandiose narratives.

For global enterprises—especially those balancing regulation, safety, and public trust—the Dutch AI approach offers a template for responsible, scalable transformation. Rather than chasing hype, these companies demonstrate how AI becomes valuable: not through magic, but through disciplined strategy, organisational clarity, and respect for the people who must ultimately use the tools.