Communication careers: skills to develop with AI
The evolution of the communication professions is a matter of tasks and responsibilities. Being able to generate a text or image becomes a capacity among others: the professional must above all know how to formulate a problem, judge a proposal and explain the decisions taken for the brand.
Article prepared with artificial intelligence assistance. The sources and method are indicated in the text.
Train on an observable task and deliverable, then assess autonomy, judgement quality and ability to correct.
Distinguish the exposure of a task and the disappearance of a profession
In its 2025 update, the International Labour Organisation is studying the exposure of occupations to general artificial intelligence from the tasks that make up them, and stresses that the transformation of work is a more likely outcome than the complete replacement of most of the jobs exposed. This global analysis does not predict the future of a specific job in an agency.
Strategy: formulate a problem before producing an answer
Teams need people who can link an observation to a business issue and then to a communication idea. Reading a study, conducting an interview, identifying a contradiction in a brief and choosing a priority remain central skills. A tool can help structure elements; it does not automatically know the client context.
Creation: Choosing and defending leadership
The creative role is not just a question of asking for an image or slogan, but also a question of recognizing a relevant track, building a visual system, maintaining an intention in adaptations and explaining its arbitrations. Learning to touch, compare and document becomes as important as producing a first jet.
Verification: installing a common discipline
A team must know how to find a source, check a quote, check an offer and identify a misleading image. These tasks should not be carried over to the last person who opens the file. Plan a responsible person and a check time for each deliverable.
The measure: linking work to outcome
A good scoreboard helps answer a decision question. A single curve does not show that a campaign has caused sales to change.
A four-step progress plan
- Mapping the team's recurring tasks and difficulties.
- Choose a limited use case with a reference before testing.
- Form a trade and technical pair; fix a human validation.
- Evaluate quality, time, costs and errors, and decide on the next steps.
This plan is used to test a new way of working. It does not justify a promise of universal productivity or an automatic decision on the future of a person.
Build a competency plan by task
Describe the activities actually carried out: research, recommendation, writing, adaptation, control and restitution. For each, specify what can be assisted, the necessary data and the level of responsibility. This mapping avoids reducing training to a list of tools. It also shows the cross-cutting skills that link the functions: The following are the main features of the project:
An exercise that reveals the level of mastery
The right result is not only an elegant presentation: it must point out what cannot be confirmed and propose a way to resolve the issue. The exercise evaluates professional judgment. He does not claim to measure an entire profession or decide on the future of a person from an automated score.
Preserving transmission to junior profiles
A team that automates all the first tasks must think about how the less experienced will learn to recognize a good job. Comparing versions, explaining corrections and participating in journals allows for the transmission of criteria. The time saved can be reinvested in this learning. Without this intention, an organization may gain immediate production while weakening its future capacity to judge.
Assessing a progression without overpromising it
Monitoring can observe the quality of deliverables, the errors identified, the necessary repeats and the ability to explain a choice. The context of exercises must be kept and the limits of evaluation recognized. The exposure of a task to artificial intelligence, as studied by the ILO, is not an individual substitute prediction; Our training recommendations should not be read as such a forecast.
Source and method
International Labour Organization, Generative AI and jobs: A 2025 update, 20 May 2025The competency grid presented here is an editorial analysis applied to the communication professions.
Continue with a practical method
Suggested routes from the story to briefing, production or measurement. These features are separate from the original publication.