There is a phrase that almost always prompts a thoughtful silence in executive forums: companies want and need to use AI. However, the greatest misconception of this era is to think that adopting personal AI is the same as transforming a company with AI.
The technology is the same. The context is not. These are two distinct revolutions coexisting within the same organisation: yours, as a professional saving time, and that of your company, which must reinvent how it creates value. Confusing the two is the most costly — and most subtle — mistake currently being made in many organisations.
This article is about the second one. The difficult one.
Adopting AI or transforming the business: the difference that will determine business success
When an individual starts using AI, what changes is their personal productivity, and this is reflected in the business. They adopt a tool, go through a personal learning curve and measure the return in terms of hours saved. The decision is straightforward: it requires a licence, training and time invested in adopting this technology. The risk lies in not using it and not adopting it. The timeframe: weeks.
When a company decides to transform itself using AI, none of that applies. It is not a matter of adopting a technology: the operating model is redesigned. The learning curve is not individual, but organisational. The return is not measured in hours saved, but in new value models. The decision is not about a licence: it is a vision sustained over time and an investment that is not diluted in the short term. The risk is no longer in not doing it, but in doing it without truly transforming. And the timeframe is measured in years.
The first act of leadership is not to confuse these two discussions. They have different agendas, different budgets, different teams and, above all, different leadership styles. When the same agenda item discusses ‘let’s give the team the freedom to act, let’s increase adoption’ and ‘how do we reinvent operations with AI’, neither makes any progress.
AI is not just another technology project: why this transformation is different
It is worth debunking a convenient analogy. This is not about adopting a new technology and, therefore, it might seem that it is solely the responsibility of the technology department.
The ERPs of the 1990s standardised back-office processes. The web and mobile technologies of the 2000s opened up new customer channels. The cloud of the following decade gave us elastic infrastructure. All those waves brought about technological change and all could be delegated, to a large extent, to a technical department.
AI does not affect just the technological layer. It affects processes, data, people, decisions and leadership. It changes how we work, how we make decisions and how we lead. And that is why it cannot be simplified as just another technological change. Any leader who signs off on the AI budget and returns to their meetings expecting ‘the IT department to sort it out’ has already lost.
The three disruptive shifts that are forcing organisations to be redesigned
Three disruptive shifts explain why this time is different.
The tool reasons. Traditional software did exactly what you asked it to do. AI makes suggestions, reasons and sometimes gets it wrong. The system is no longer deterministic, and the organisation has to learn to live with probability. This is nothing new when it comes to people; it is radically new when it comes to technology.
Adoption is personal. There is no single ‘go-live’ moment, no date when the system is switched on and the project ends. Each person uses AI differently, and the transformation takes place through thousands of daily micro-decisions that no one approves in a committee.
The value does not lie in the technology. It lies in the process you redesign around it. Buying licences is not the same as transforming. The real investment lies in changing the way we work, and that inevitably leads to a new leadership model.
When the foundations shift, new technology alone is not enough. A new operating model is required.
The five essential dimensions of an AI operating model
Here lies the crux of the matter. The AI operating model rests on five levers, and the rule is an uncomfortable one: you must move them all at the same time. If you make progress on one and leave the others behind, the system breaks down. Transformation is either simultaneous or it isn’t.
1. Strategy and purpose: defining where AI creates value
We need an explicit AI thesis, in a single sentence: where it creates real value for us — efficiency, experience or new business models. From there, a dynamic portfolio of use cases prioritised by impact and difficulty, not a collection of never-ending pilots that never scale. And metrics that go beyond ROI per case: the percentage of processes redesigned, of people actively involved, and of hours freed up.
The litmus test is simple: if AI isn’t discussed at every top-level committee meeting, it’s not a strategy—it’s a project. Without a thesis, pilot projects don’t accumulate: they fizzle out and we never transform.
2. Data, identity, privacy and technology: the true accelerator of enterprise AI
Without a data catalogue, clear permissions and a minimum standard of quality, enterprise AI won’t scale. The architecture must be modular; it need not be centralised — a layer of in-house and third-party models, an orchestration layer and an application layer — to enable rapid change and avoid lock-in with a single supplier.
And everything must be governed and traceable by design: who requested what, using which model and with which data. If it cannot be audited, it cannot be trusted.
The speed at which we transform ourselves with AI depends on data quality, on having a traceable identity across domains, on managing privacy by design, and on governance capabilities; not the other way round.
3. Processes and work: redesigning from scratch to capture the value of AI
The most common mistake is to automate the legacy process. The right question is not ‘how can I do what I already do faster?’, but ‘how would I do it today if the company were founded with AI?’.
We need to map the work — the tasks repeated a thousand times a day — not the boxes on the organisational chart. And we need to design the human + AI architecture as standard: where the person decides, where the machine makes suggestions, and where traceability lies.
If the process remains the same and only the tool changes, there is no transformation.
4. People and culture: the factor that most influences the success of the transformation
Applied AI is for everyone, not just an elite team. Training starts with 100 per cent of the workforce and is learnt by doing: through practice, internal communities and real-world challenges. The course is the practice itself; theory alone is not enough.
Most roles will evolve and change significantly, although the tasks and responsibilities associated with them will differ. It is the leader’s sole responsibility to drive that change before it happens of its own accord.
And none of this works without explicit permission to experiment and a clear protocol for failure. Adoption is not decreed: it is cultivated.
5. Governance and risk: trust to accelerate AI adoption
Written principles are one thing; the decisions teams make every morning are another.
We need an operational AI committee with real power: one that assesses cases, puts the brakes on those that need stopping, and gives the green light to those that add value.
Compliance — business impact, regulation, data protection, intellectual property and sustainability — must not act as a brake, but rather as the safety net that allows us to accelerate with confidence.
And transparency with employees: what data is used, what is automated and what control remains with the individual.
Without trust, AI slows itself down. Governance is not a brake: it is an accelerator. And if we want real operational impact, it must be built in from the design stage.
The new leadership required by the age of artificial intelligence
Here lies the most profound challenge for organisations, because it strikes at their very core: culture.
When the tool reasons, makes proposals and, in certain contexts, can carry out actions within defined limits, the traditional leadership model is re-evaluated.
Managing tasks gives way to designing human + AI systems. Demanding certainty gives way to living with probability. And ‘knowing more than the team’ gives way to ‘deciding better with the team and with AI’.
Implementing AI in a company is, at its core, implementing a new leadership model.
Five capabilities define the leader of this era:
- Judgement regarding certainty. Making decisions based on partial information and probabilistic reasoning, without waiting for the exact data that will never arrive.
- Designing human-AI systems. Allocating tasks between people and machines with intention, not by accident.
- Honest technical curiosity. Understanding enough to ask the right questions. Not to plan in a deterministic way; but to exercise sound judgement.
- Caring for the team during the transition. Making the change in roles explicit and addressing fears. Transformation depends on psychological safety.
- Ultimate, non-delegated responsibility. AI makes proposals; decision-making and accountability remain human.
The time to act is now: the decision that will shape the future of organisations
AI is not here to make minor improvements to what already exists. It is here to challenge the way we work, the way we make decisions and the way we lead. And for that reason, transforming a business with AI is not a technological project: it is a project of business reinvention.
In this decade, the organisations that will make a difference will not be those that buy the most licences or launch the most pilot schemes, but those that dare to redesign themselves before the rest.
Because the real competitive advantage will not lie in accessing the technology, but in having the vision, the courage and the ability to execute in order to turn it into a new way of doing business.
Transforming a business with AI means much more than simply incorporating new tools. It means building capabilities to compete in an environment where technology, talent and data have become strategic factors. In line with Telefónica’s vision of “becoming the best gateway for citizens to digital technologies” and “contributing to Europe’s technological sovereignty”, the real challenge lies in leading a transformation that combines innovation, trust and the ability to execute. Because the future will not belong to those who adopt AI first, but to those who are capable of reinventing their organisation around
