AI has driven the cost of content to zero. Here's what still has value
When producing costs nothing, producing stops being a service. What remains payable in corporate communication.
For twenty years the communication market rested on a simple asymmetry: writing well was hard and expensive, so those who could do it could charge for it. That asymmetry is gone. A generative model produces in nine seconds a text that in 2019 would have taken half a day and three hundred euros.
The hasty conclusion is that professional communication is finished. The correct conclusion is different: the point where value is created has moved.
What AI does better than a professional
Research and document synthesis. Media monitoring and sentiment analysis at volumes no human would read. Building and cleaning media lists. First drafts and variant generation. Preparing Q&A and simulating objections. On all of this there's no contest, and delegating is the right call.
What AI does not do
The thesis. A language model can write any position, and precisely for that reason it chooses none. Deciding what a company defends publicly — and what it's willing to lose defending it — is a risk decision, not a language one.
Judgement on reputational risk. Knowing that a given paragraph, read by a given counterpart, at a given moment in the market, will produce an unpleasant phone call is not a linguistic skill. It's compressed experience.
Relationships. A journalist doesn't respond to content. They respond to a person they trust because that person hasn't wasted their time before.
The choice to stay silent. Perhaps the most underrated competence. Generative systems are designed to produce; value often lies in not publishing.
The practical consequence for budgets
If you're still paying someone by volume — so many pieces a month, so many press releases — you're buying the part that became free and giving away the part that matters. It's time to rewrite the contract: fewer deliverables, more decisions; less output, more time from someone who knows where not to go.
There's also a market effect worth anticipating. When everyone produces more, scarcity shifts to attention. In an ecosystem saturated with correct, indistinguishable text, the only thing that gets noticed is a recognisable position, repeated and attributable to a person.
AI hasn't made communication cheaper. It has made it more demanding: the one thing you can't automate is having something to say.
