Where AI (truly) creates value

A few weeks ago we took part in AI Festival, one of the contexts where the debate around Artificial Intelligence is particularly active. Much was said about new features, increasingly powerful tools, and promises that seem to open up endless possibilities.

It is precisely on occasions like this, however, that a less immediate yet decisive question emerges. A question we often encounter when working on digital products and services: not so much what we can do with AI, but where it truly makes sense to use it.

Shifting the focus from the technology’s possibilities to organizational processes radically changes the perspective. It means asking where AI can help solve concrete problems, improve existing experiences, and be meaningfully adopted in everyday work.

If there is one thing we have learned in over a decade of digital transformation projects, it is that when a project fails to achieve the expected results, the issue is rarely the technology itself. More often, the technology does not align with the context and organizational processes, or decisions about features and content have been made at the wrong stage of the process, or at the wrong user touchpoint.

Very often, this stems from a solution-first perspective (e.g., “let’s change the intranet platform,” “let’s rebuild the app”), where the problem-framing phase is skipped. In scenarios that involve AI, this gap becomes significantly more risky, given the speed and autonomy with which such systems operate.

Starting from processes to understand where AI can create real value

Before imagining new features or assistants, it is useful to pause and examine what already exists: processes, services, workflows. Starting from processes helps identify where work actually gets stuck - in repetitive and time-consuming steps, bottlenecks, or moments in the experience that create friction and tension.

It is precisely in these points that AI can make sense as a component of a broader solution. But it is also here that it becomes clear where introducing it at all costs does not make sense.

This is an approach we have long applied at Tangible when working on complex digital services and products, and one we have already explored in previous Thinking articles focused on reading processes as a design lever.

Workshops such as an AI Opportunity Map are designed precisely to support decision-making and make explicit the choices made throughout the process. They help clarify what is worth exploring and what, instead, would add complexity without generating value.

Not all AI solutions are the same

When observing a process through this lens, three main directions typically emerge in which AI components can intervene:

• They can act invisibly, automating repetitive parts of the work
• They can support people in decision-making, by suggesting, summarizing, or highlighting relevant information
• Or they can become an explicit interface, a direct point of contact between system and user.

None of these options is inherently better than the others. Each implies different choices in terms of experience, responsibility, and adoption.

Deciding which scenario makes sense in a given context is already an act of design. Often, it is this choice, more than the technology itself, that determines whether an AI-powered solution will actually be used.

Adoption as a signal of success

As with digital transformation projects, when discussing AI-powered solutions there is a risk of confusing value with the mere presence of technology. A feature can be technically sophisticated, well designed, even appreciated during a demo — and yet never truly enter people’s daily practices.

In complex contexts - such as structured companies, public organizations, and articulated digital ecosystems - adoption is not an automatic consequence. It is a design choice.
Much depends on where the solution is placed within the process, how well it integrates with existing tools and habits, and whether it is perceived as genuine support rather than an additional cognitive load.

For this reason, when working on solutions that integrate AI, adoption should become one of the primary signals of success. What matters is that the solution is used and proves useful when it is needed.
Taking it one step further: once adopted, what business value does the solution actually shift?

Choosing what to observe in order to understand whether we are on the right path is particularly important in projects with an experimental component, such as those involving AI. In fact, it is a decisive design choice that shapes the direction of the project itself and therefore must be intentional and conscious.

Looking at adoption in these terms also helps refocus the role of design as the space where decisions are made that determine whether a solution will have real impact or remain theoretical.

Choosing is already designing

When viewing the integration of Artificial Intelligence from this perspective, it becomes clear that its value does not emerge automatically with the introduction of a new technology. It emerges from the choices made beforehand, starting with where to intervene and what role AI should play within the process.

Designing intelligent solutions today means taking responsibility for these decisions, accepting that some opportunities should be left aside and that not every problem requires AI as a technological answer.

It is within this space - made of conscious and often less visible choices - that AI stops being an abstract promise and begins to become an integral part of systems that people actually use. And where the work of design continues to be, first and foremost, a work of direction.