If AI doesn’t transform, it’s “business as usual,” not true intelligence

Companies are investing more and more in artificial intelligence. Projects and experiments are multiplying, yet in most cases the impact remains surprisingly limited: a few efficiency gains, some improved decisions, but nothing that truly changes the way organizations operate.

The first of two studies conducted by SDA Bocconi with support from Google starts from the observation that two companies can invest similar amounts in AI and achieve completely different results. And the reason is not the technology itself, but the way it is introduced, governed, and integrated into the organization.

In many cases, AI remains confined to individual projects disconnected from one another. It improves certain activities without affecting core processes, strategic decisions, or the business model. In other words, even when adopted, it does not transform.

Beyond the local level

The debate around artificial intelligence has focused mainly on its ability to automate tasks, analyze data, and support decisions. But this perspective leaves one question unanswered: why, despite all this potential, does AI’s real impact within companies often remain modest?

The literature has already identified some clues. AI generates value, but mainly at the local level: within a function, a team, or a process. The problem is that this value rarely translates into a broader competitive advantage.

This is where the study’s questions emerge:

• How does AI create value within organizations?
• Where does this value emerge?
• Why do some initiatives scale while others do not?
• How do companies move from many isolated projects to true transformation?

A two-dimensional map

The researchers conducted an in-depth analysis of five large organizations operating across different sectors, from manufacturing to finance, tourism, and advanced engineering. They also carried out interviews with executives and AI leaders.

The result was a framework that interprets AI along two dimensions.

The first concerns how AI creates value:

• by automating activities,
• by supporting decisions,
• or by radically changing the way work is done.

The second concerns where this value emerges:

• at the local level,
• across multiple functions,
• or at ecosystem scale, involving partners and customers.

Crossing these two dimensions produces a map of possible strategies. Most companies concentrate in the “simplest” area of the map: automation and enhancement at the local level. This is where the quickest and least risky results can be found, but it is also where expanding the impact across the organization becomes difficult.

The most transformative applications — those that redesign processes, operating models, or entire value chains — are found elsewhere and require far more: integration, governance, shared data, and deep organizational change.

Consistency, data, and organization

For companies that truly want to make a difference, there is no universally “best” configuration, also because not every organization needs to pursue radical transformation. What really matters is consistency. Organizations that achieve results are those in which strategy, data, governance, and use cases are aligned.

Without this consistency, the most common risk is the accumulation of projects. When every function develops its own use cases, the company ends up building a disordered mosaic of solutions that are difficult to integrate and maintain.

Another factor limiting AI scalability is data. Without a solid, shared, and well-governed data foundation, AI remains confined to local initiatives.

The organizational dimension emerges in one final reflection: if the introduction of AI does not change roles, responsibilities, and decision-making processes, its impact will remain superficial.

Starting from the foundations

The study suggests that managers should avoid thinking about AI as a race toward a finish line. There is no single destination that fits everyone. Instead, there are strategic choices, with trade-offs between impact, complexity, and risk. Managers should also do everything possible to avoid the “use case trap,” which creates the illusion that doing more means transforming more. In reality, without integration, more projects simply mean more complexity.

Corporate investments should start from the foundations: data, architectures, and skills. This is where the real possibility of scaling lies.

Ultimately, AI creates value only when it forces the organization to change, becoming a catalyst for realigning strategy, processes, and people.

Nico Abbatemarco, Gianluca Salviotti, Carmelo Cennamo. AI transformation logics.