La pregunta que todos los líderes se están haciendo — y por qué la mayoría la está haciendo mal.
Hace poco, un colega facilitador de cambios culturales hizo una pregunta que representa lo que miles de líderes se están preguntando hoy: ¿Podría la IA ayudar a transformar la cultura organizacional? ¿La reflexión crítica, el pensamiento sistémico, la colaboración?
Es una pregunta honesta. Y tiene un supuesto invisible que vale la pena nombrar: asume que la IA es un agente que actúa sobre la cultura. Que algo externo puede entrar en una organización y transformarla.
Ese supuesto tiene nombre. René Descartes lo instaló hace cuatro siglos: la idea de que el mundo es un mecanismo que puede ser intervenido desde afuera, que las causas vienen de fuera hacia adentro, que si introduces el estímulo correcto obtienes el resultado esperado.
Es un supuesto poderoso. Y en este caso, es incorrecto.
La cultura organizacional no es lo que está escrito en la pared. No son los valores declarados ni la misión del sitio web. El sociólogo Darío Rodríguez la definió con una precisión que no he encontrado igualada en veinticinco años de consultoría:
"La cultura es el conjunto de premisas básicas desde las cuales se construye el decidir organizacional."
Premisas. No valores. Premisas operativas — los supuestos que todos en la organización comparten sin haber acordado nunca compartirlos, que gobiernan lo que se decide y lo que no se decide, lo que se puede cuestionar y lo que es intocable.
Son invisibles porque son fundamentales. Son el agua en que todos nadan. Y eso las hace extraordinariamente difíciles de transformar — no porque la organización no quiera cambiar, sino porque no puede ver aquello desde lo cual está viendo.
Desde la biología del conocer de Humberto Maturana, los sistemas vivos son estructuralmente determinados: lo que le ocurre a un sistema ante cualquier perturbación externa no está determinado por esa perturbación, sino por la estructura del sistema que la recibe.
En términos simples: nadie puede cambiar a nadie. Solo se puede perturbar. Y lo que ocurre después depende de la estructura del sistema perturbado, no de la calidad de la intervención.
Perturbar los circuitos comunicativos de una organización. Introducir distinciones nuevas. Acelerar el acceso a información. Hacer visible lo que los patrones culturales invisibilizan. Crear condiciones para que emerja algo nuevo.
Transformar la cultura. Las culturas las transforman las comunicaciones que los sistemas psíquicos — las personas — producen entre sí. La IA puede ser una condición de esa transformación. No es su agente.
Confundir esos dos niveles no es un error semántico. Es un error de diseño que produce intervenciones que instalan tecnología sin transformar nada — y dejan a las organizaciones más confundidas que antes.
Sirve — y puede servir extraordinariamente bien — cuando se usa para lo que genuinamente puede hacer: perturbar de manera generativa.
He observado que las organizaciones que integran IA de manera que produce valor cultural real comparten una característica: no usan la IA para que les diga qué hacer. La usan para verse a sí mismas de una manera que antes no podían.
Un sistema de inteligencia artificial no comparte la historia de la organización. No tiene las lealtades, los miedos, las relaciones de poder que estructuran lo que se puede decir y lo que no. Esa distancia — bien utilizada — puede revelar premisas que la organización no sabía que tenía.
La IA no transforma la cultura. Puede crear las condiciones para que la organización se perturbe a sí misma de manera que produzca transformación.
Esa diferencia no es filosófica. Es práctica. Determina cómo se diseña la integración, quién asume la responsabilidad del proceso, y qué se puede esperar — y qué no — de la tecnología.
Este texto es un adelanto de un libro en proceso. Su pregunta central no es qué puede hacer la IA — es qué ocurre cuando algo que no es humano entra en el dominio donde lo humano se constituye: el lenguaje, el sentido, la comunicación.
El libro cruza tres marcos teóricos que raramente se encuentran en la misma conversación sobre IA: la biología del conocer de Maturana y Varela, la teoría de sistemas sociales de Niklas Luhmann, y el concepto de comunicación artificial de Elena Esposito. Y los cruza con veinticinco años de trabajo en organizaciones reales.
No es un libro sobre tecnología. Es un libro sobre lo que la tecnología nos obliga a preguntarnos sobre nosotros mismos.
— English version follows —
The question every leader is asking — and why most of them are asking it wrong.
Recently, a colleague — a cultural change facilitator — asked a question that represents what thousands of leaders are wondering today: Could AI help transform organizational culture? Critical reflection, systemic thinking, collaboration?
It's an honest question. And it carries an invisible assumption worth naming: it assumes that AI is an agent that acts upon culture. That something external can enter an organization and transform it.
That assumption has a name. René Descartes installed it four centuries ago: the idea that the world is a mechanism that can be intervened from the outside, that causes flow from outside inward, that if you introduce the right stimulus you get the expected result.
It's a powerful assumption. And in this case, it's wrong.
Organizational culture is not what's written on the wall. It's not the declared values or the mission statement on the website. Sociologist Darío Rodríguez defined it with a precision I haven't found matched in twenty-five years of consulting:
"Culture is the set of basic premises from which organizational decision-making is constructed."
Premises. Not values. Operative premises — the assumptions that everyone in the organization shares without ever having agreed to share them, which govern what gets decided and what doesn't, what can be questioned and what is untouchable.
They are invisible precisely because they are fundamental. They are the water everyone swims in. And that makes them extraordinarily difficult to transform — not because the organization doesn't want to change, but because it cannot see that from which it is seeing.
From Humberto Maturana's biology of knowing, living systems are structurally determined: what happens to a system in response to any external perturbation is not determined by that perturbation, but by the structure of the system receiving it.
Simply put: no one can change anyone. You can only perturb. What happens next depends on the structure of the perturbed system, not on the quality of the intervention.
Perturb an organization's communicative circuits. Introduce new distinctions. Accelerate access to information. Make visible what cultural patterns render invisible. Create conditions for something new to emerge.
Transform culture. Cultures are transformed by the communications that psychic systems — people — produce among themselves. AI can be a condition for that transformation. It is not its agent.
Confusing these two levels is not a semantic error. It's a design error that produces interventions that install technology without transforming anything — and leave organizations more confused than before.
It's useful — and can be extraordinarily useful — when used for what it genuinely can do: perturb generatively.
I've observed that organizations that integrate AI in ways that produce real cultural value share one characteristic: they don't use AI to tell them what to do. They use it to see themselves in ways they couldn't see themselves before.
An artificial intelligence system doesn't share the organization's history. It doesn't have the loyalties, the fears, the power relations that structure what can be said and what cannot. That distance — well used — can reveal premises the organization didn't know it had.
AI doesn't transform culture. It can create the conditions for an organization to perturb itself in ways that produce transformation.
That difference is not philosophical. It's practical. It determines how integration is designed, who takes responsibility for the process, and what can — and cannot — be expected from the technology.
This text is a preview of a book in progress. Its central question is not what AI can do — it's what happens when something non-human enters the domain where the human is constituted: language, meaning, communication.
The book crosses three theoretical frameworks rarely found in the same conversation about AI: Maturana and Varela's biology of knowing, Niklas Luhmann's theory of social systems, and Elena Esposito's concept of artificial communication. And it crosses them with twenty-five years of work in real organizations.
It is not a book about technology. It is a book about what technology forces us to ask about ourselves.
Todo libro honesto termina con una pregunta más profunda que la que abrió. Este no es la excepción.
Cristián Cordero Placencia — Procesus — Isonauta