Introduction to the series ‘AI in action: 9 steps to move from use to value creation’

1. A historic moment, comparable to major technological revolutions

At certain periods in economic and technological history, innovations emerge that do more than simply improve what already exists. They reconfigure the rules of the game, disrupt value chains and redefine competition.

In 1994, the advent of the Netscape browser marked a turning point. The internet, which until then had been confined to laboratories and universities, opened up to the general public. Within a few years, companies that understood its potential – Amazon, eBay and Yahoo in their early days – established themselves as dominant players. Others, though well established, failed to make the leap.

In 2007, the arrival of the iPhone transformed the telephone into a universal platform. Apple didn't just launch a new product: it created an entire ecosystem, where the smartphone became the access point to banking, commerce, social networks and work. The network effect was such that entire sectors of the economy – transport with Uber, accommodation with Airbnb, distribution with Deliveroo – were born out of this reinvention.

In the 2010s, cloud computing shifted the centre of gravity of IT. Amazon Web Services, Microsoft Azure and Google Cloud didn't just sell servers: they changed the way companies think about their infrastructure, moving from a heavy CAPEX model to a flexible OPEX model.

Today, in 2025, we are experiencing a similar moment with artificial intelligence, and more specifically with generative AI and agentic AI.

This is not a passing fad, but a systemic disruption. Companies that recognise this ‘Netscape moment’ and respond to it will not just improve their processes: they will invent new ways of working, create new services and, in some cases, open up entire markets.

This transformation is already evident in several ways:

• DuPont is using generative AI to accelerate the discovery of new polymers. Whereas research cycles could take several years, AI reduces this time to a few months, increasing both efficiency and innovation capacity.
• Morgan Stanley has deployed an internal AI co-pilot that aggregates thousands of research and compliance documents to assist its financial advisers. The result is more personalised advice for clients and time savings in information retrieval.
• In healthcare, Sanofi is experimenting with generative AI to analyse clinical trials and improve protocol design, accelerating the time to market for drugs.
• In the energy sector, BP is combining generative AI and digital twins to optimise its predictive maintenance operations.

These examples show that AI is not just a productivity tool: it is a catalyst for structural transformation.

2. The widening gap: users vs. value creators

The latest data reveals a striking observation: all companies report ‘experimenting’ with AI, but very few derive measurable and sustainable value from it.

Stanford's AI Index 2025 indicates that 72% of large companies already use AI in some way. However, only 14% of them report that these deployments have a direct and significant impact on their financial performance.

Why such a gap?

The difference lies in two strategic approaches:

• AI users: they consume the tools available on the market, often in an opportunistic manner. They install a chatbot, test an integration with Office, experiment with a few POCs. This can generate one-off gains (a little productivity, marginal savings), but nothing that really sets them apart. These organisations run the risk of ‘AI washing’: they talk a lot about AI, but without any tangible impact on their competitiveness.
• AI value creators: they integrate AI into the heart of their strategy. They don't just adopt a tool: they build internal platforms, connect their proprietary data and reconfigure their workflows. AI becomes a central part of their business model.

Let's take an example from industry: Siemens. Rather than limiting itself to deploying a few ad hoc solutions, Siemens has integrated AI into its digital twin platform, MindSphere. As a result, AI is not an additional layer, but a catalyst that optimises production, reduces energy costs and offers industrial customers unprecedented simulation and predictive maintenance capabilities. Siemens has become a ‘value creator through AI’.

Conversely, many industrial players are still content to integrate a ‘generic co-pilot’ into their office tools. They improve the efficiency of certain administrative tasks, but their competitive advantage remains unchanged. They remain ‘users’ rather than value creators.

3. Why a 9-step series?

This series of newsletters has been designed to guide this transition.

The objective is simple: to provide a clear strategic framework, divided into nine steps, to help executives and managers move from superficial use to sustainable value creation.

These nine steps follow a logical progression:

1. Understand the historical moment — the ‘Netscape moment’ of AI.
2. Move from user to value creator.
3. Know how to convince an executive committee with solid arguments.
4. Identify use cases that truly create value.
5. Integrate trust and governance as conditions for survival.
6. Develop internal skills to support the transformation.
7. Choose the right technology models (SLM, LLM, MoE).
8. Leverage data as a strategic differentiator.
9. Anticipate the future of generative computing.

Each week, a newsletter will explore one of these themes.

The ambition: to go beyond technological fascination and provide strategic benchmarks for decision-makers.

4. The major cross-cutting issues of 2024-2025

Against the backdrop of this series, several structural trends are already redefining the landscape.

Regulation

In 2024, Europe adopted the AI Act, the world's first systemic regulation on AI. Its obligations will be phased in until 2027, but some provisions will apply as early as 2025: prohibition of practices deemed unacceptable, transparency requirements for generative models, and strict compliance for high-risk systems.

In the United States, the NIST GenAI profile provides an operational framework for measuring, auditing, and securing generative AI systems.

For businesses, the issue is no longer just technical: it is legal, ethical, and reputational. Tomorrow's leaders will be those who have integrated regulation into their governance today.

Data as capital

Less than 1% of corporate data is included in large public models. Yet it is this proprietary data (contracts, production histories, customer data) that holds most of the value.

Companies that are able to secure, manage and leverage their data will be the ones that turn AI into a sustainable competitive advantage.

Skills development

According to the World Economic Forum 2025, 40% of current skills will need to be reinvented by 2030. The key is not only to hire AI experts, but also to provide in-depth training to existing employees.

AI will not replace humans, but those who know how to use it will replace those who do not.

Infrastructure

The technological paradigm is evolving: small language models (SLMs) are developing, offering targeted performance at lower cost and with reduced latency. Mixture of Experts (MoE) architectures combine several specialised models. Hardware accelerators (NVIDIA Blackwell, IBM NorthPole) reduce energy costs and pave the way for real-time AI.

Companies must now consider a multi-model and multi-cloud strategy.

Sustainability

Finally, the energy consumption of large models has become a central issue. Companies will need to find a balance between performance and sustainability. AI will also be judged on its carbon footprint and alignment with ESG objectives.

5. Invitation to the reader

This series of newsletters is not intended to be yet another technical guide.

It aims to create a strategic and narrative framework, supported by concrete examples and recent data, to enable executives to consider AI as a lever for transformation.

Each week, a new episode will bring a different perspective, but all will have a common thread: moving from use to value creation.

The real question is not whether we will adopt AI.

It is whether we will remain mere consumers... or become creators of value.

6. Forward-looking conclusion: what does the next five years hold in store for us?

By 2030, the dividing lines will be clear:

• companies that have integrated AI into their processes will have gained in productivity, agility and innovation capacity.
• The others will remain dependent on model providers, unable to differentiate themselves.

We can already anticipate three major developments:

1. Agentic AI: autonomous systems capable of achieving complex objectives by cooperating with each other.
2. Generative computing: a new computing paradigm, where AI will no longer be just a tool, but a mode of computation in its own right.
3. The convergence of AI, data and regulation: AI will become inseparable from data governance and regulatory requirements.

In this changing world, leaders have a historic responsibility: to prepare their organisations not only to survive, but also to thrive.