When AI silently reshuffles the cards of “who does what”

When I support companies in integrating AI, I always observe the same thing: the tool arrives officially, but what it transforms first is not processes, it is professional responsibilities.

This shift that I observe often happens without a formal decision, without a collective decision, and yet it profoundly changes the way work is distributed.

When work changes before roles do

In the field, I see very few obstacles related to the tools themselves. On the other hand, I regularly observe silent shifts in the day-to-day reality of roles.

Some tasks disappear, others move to different people. Certain missions lose part of their purpose, without being redefined.

Very quickly, questions emerge among employees. They are rarely expressed out loud. They often remain silent, but they are very present:

“What is expected of me now? Where is my value today?”

These questions arise well before any structured HR reflection.

When certain missions disappear… and no one talks about it

In my support work, I have seen tasks such as proofreading, rephrasing, verification, or preparation gradually taken over by AI. Teams see it, and they understand it very quickly.

In many organizations, I observe that words are rarely put on this, and that there are still far too few spaces to talk about what work is actually becoming.

The observations remain individual and, when they remain that way, they become anxiety-inducing. Everyone adjusts on their own, with their own interpretation.

When a departure becomes an opportunity to rethink a role

In a company I was supporting, the assistant left her position. The executive already knew he was not satisfied with the outcome. The work was done, but the added value had remained limited.

The temptation would have been simple: recruit a new person using the same job description.

I alerted him very clearly. With AI already integrated into your practices, I told him that starting again with a copy-paste made no sense.

It was an opportunity to evolve the role, exactly as he wanted it. If he wanted real support, then it was necessary to delegate everything repetitive and time-consuming to AI, and to redefine this role with more responsibilities, coordination, and genuine support.

Result: the recruitment changed in nature, with a role that was more strategic, more structuring, and useful both to the executive and to the company.

Always the same people absorbing the transition

Another phenomenon comes up frequently in what I observe. It is always the same profiles who take charge of the evolution.

Those who quickly understand AI, grasp its logic, and spontaneously structure its uses. They absorb new responsibilities, without this being formalized, and very often without clear recognition of the shift in their role.

Meanwhile, others remain confined to the old perimeter, not due to a lack of willingness, but due to a lack of framework to evolve.

This is not an individual problem. It is a problem of unmanaged redistribution of work.

La question, c’est de savoir si ces déplacements sont discutés, arbitrés et assumés dans votre organisation, ou s’ils se font en silence..

Redéfinir les responsabilités est un enjeu de cadrage

Repenser les responsabilités professionnelles devient indispensable, même si cela touche à des zones sensibles : 

• les habitudes, 
• les fiches de poste, 
• la représentation de la valeur, 
• le temps travaillé, le salaire.

However, if you allow these evolutions to happen implicitly, this creates far more imbalances than opening the discussion.

A role can evolve, a mission can transform; what is unavoidable is the implementation of a structured framework to support these changes progressively.

Letting AI redraw roles in place of management is a major risk.

Poser un cadre pour travailler à partir du réel

This work does not rest solely on the executive. It requires team managers capable of observing what is actually changing on the ground. It also requires spaces to name these evolutions without taboo.

When roles need to be clarified, this is not about threatening people. On the contrary, it is about securing the collective.

When this framework is in place, responsibilities gain in added value. Missions become more interesting. Work regains meaning.

AI does not remove work. It forces us to redefine it consciously.

C’est précisément à ce niveau que j’interviens : amener les cadres et responsables d’équipe à poser un cadre clair, éclairer les usages réels de l’IA et faire évoluer la manière de travailler sans laisser les équipes absorber seules ces transformations.

If this article has brought clarity, d’autres textes prolongent cette réflexion other pieces extend this reflection on AI practical uses, and the transformation of professional practices. 

About the author

Micheline Boutrin Deroire

Fondatrice de PAS À PAS DIGITAL, consultante stratégique en gouvernance et intégration des usages professionnels de l’intelligence artificielle.

Elle accompagne les cadres et responsables d’équipe sur le cadrage des usages, les décisions qu’ils impliquent et les responsabilités professionnelles associées.

Son approche relie quatre axes — IA, gouvernance, cadrage des usages et responsabilité professionnelle — pour structurer les pratiques sans les dénaturer.

Ses travaux s’ancrent dans le terrain, les situations réelles et les enjeux organisationnels des entreprises.

👉🏼 Continue the reflection or start a conversation

[email protected] | +596 696 37 66 90

(Distinctions 2024 : Best Innovation & Strategic Change Consultancy Leader – Western Europe * Digital Transformation Expert of the Year – France)