What this episode is about
Luis Miguel Barral is a geographer. He studies human groups, not individuals. And when he listens to a company talk about artificial intelligence, he hears the same thing he hears when it talks about any other transformation: fear disguised as strategy.
This episode starts where the headlines end. 98% of corporate AI pilots never make it past the proof-of-concept stage. 9% of layoffs in the United States in 2025 were attributed to AI. Employees at companies that implemented AI without changing their processes are taking 350% longer to complete the same tasks. These numbers do not add up, and that contradiction is exactly the thread we pull in this episode.
Two types of company, one reality
Some companies buy an AI licence, send an internal email, and pin on the “tech company” badge. Others understand that AI is not a piece of software but a catalyst that forces you to rethink how every process, every team, and every incentive works.
Luis Miguel frames it differently: organizations that live in appearances versus organizations that live in coherence. Those that produce a hollow, cardboard-stone discourse and those that build a culture where what is said and what is done actually align.
The 350% slowdown is not a failure of AI. It is the failure of layering new technology on top of old processes without asking what lies beneath.
What AI still cannot do
Luis Miguel works with human discourse. His method consists of listening to what people say in what they say, not in what they say literally. That distinction sounds subtle, but it is enormous.
AI models feed on what is produced, on the surface of discourse. Reaching an interpretation of what lies behind what a human being says, what motivates a group, what divides a team, requires the trained judgment of an analyst. Not because AI is not intelligent, but because real listening requires emptying oneself of preconceptions, and that cannot be learned from data.
The three materials of an organization
Jorge Wasserberg, theoretical physicist and founder of the Caixa museums, proposed that the universe is made of three types of matter: inert matter (which provides stability), dynamic matter (which adapts), and cultural matter (which anticipates). Luis Miguel applies the same logic to organizations.
A company with only stability becomes rigid. One with only dynamism drifts without roots. One with only anticipation lives in the world of ideas without ever touching the ground. The balance between all three is what allows a company to survive and grow when the environment changes, which it always does.
Creativity as an advantage and as a practice
Luis Miguel’s thesis is direct: AI will replace everything that is repetitive, executable, and predictable. What it will not replace is the capacity to expand the world, to connect ideas that were not connected, to create.
And creativity can be trained. Not through methodologies or workshops, but by leaving space for boredom, for wandering, for wonder. The problem is that we have designed organizations that reward a full calendar as a signal of productivity, leaving no time or space for something different to emerge.
The news radar
MIT on AI adoption. 98% of pilots in the sample failed to scale beyond proof-of-concept. The question is not whether AI works, but whether companies are ready to change what lies beneath it.
AI washing. The CEO of OpenAI warned about companies using AI as a cover story to justify decisions that have little to do with technology. 9% of US layoffs in 2025 were attributed to AI. The suspicion is that many of those had more to do with poor management than with automation.
Junior profiles under pressure. More than 50% of layoffs in financial services are hitting profiles aged 22 to 25. We are eliminating the operational work that used to be the training ground for future senior roles. It is not clear we have thought through the consequences.