The Team You Build After AI
AI does not replace marketing roles. It collapses the buffer roles — the people who existed to move information and manage coordination. What remains is judgment, and most organisations are not structured to develop it.
Every few months, someone publishes a piece on how many marketing jobs AI will eliminate. The number changes. The framing doesn’t. It is always a substitution story: AI does X, so humans no longer need to do X.
This misses the more interesting — and more disruptive — shift.
The buffer layer
Most marketing teams have a hidden structural layer: people whose primary function is not creation or strategy, but coordination and translation. They move briefs between stakeholders. They turn research into summaries. They adapt content from one format to another. They attend meetings so the people who made decisions don’t have to attend the follow-up meetings.
This layer exists because information is expensive to move in large organisations. AI does not eliminate the need for this work. It makes it nearly free.
When coordination becomes cheap, you do not need fewer people — you need different people. Specifically, you need people who can exercise judgment at the level where coordination used to happen.
What judgment actually means
Judgment is not a personality trait. It is a capacity that develops through exposure to decisions and their consequences. A team that has outsourced its thinking to briefs, templates, and process rarely develops it.
The organisations that will build effective AI-augmented teams are not the ones asking “what can we automate?” They are the ones asking: “what decisions do we need people to own — and are we giving them the conditions to own them?”
That is a different design problem. It requires rethinking how roles are scoped, how information flows, and where accountability actually sits.
The structure question
Most marketing functions are structured for production, not for judgment. Specialisation is deep. Handoffs are frequent. The person closest to the customer insight is rarely the person with authority to act on it.
AI accelerates the production layer. If the judgment layer is thin — and in most organisations it is — acceleration just produces more of the wrong things faster.
The question worth asking is not “how do we use AI?” It is “what kind of team do we need to be?” The answer shapes everything else.