The brief came in on a Tuesday. By Thursday the copy was approved—headline, body, three social variants, email subject line, a 30-second script, and a disclaimer that legal actually liked. The client said it was the smoothest approval process they’d had in years. Nobody asked who wrote it. The answer, if they had asked, would have been: a prompt, a model, and a creative director who spent forty minutes editing and calling it their own. This is the new normal. It arrived quietly, without ceremony, and with remarkably little resistance from anyone except the people whose jobs it was replacing.
The Moment the Bar Moved and Nobody Announced It
For a while, the standard defence against AI-generated copy was quality. The outputs were serviceable but flat. They lacked voice. They could write a product description but not a sentence that made you stop scrolling. That defence is now largely gone, and the industry is still pretending it isn’t.
The copy that most brands produce—the website body text, the email nurture sequences, the social captions, the ad variants, the product descriptions, the FAQ answers—was never particularly distinguished to begin with. It was professional. It was on-brand. It was approved. And now that machine-generated text can clear the same bar, the question isn’t whether AI is good enough. It clearly is, for most of it. The question is what we were actually valuing when we paid humans to produce work of that quality, and whether that valuation was ever really about the quality.
The answer, it turns out, is complicated. We were paying for quality, yes. But we were also paying for accountability, for relationship, for the performance of craft that made clients feel like they were buying something real. The machine doesn’t perform that ritual. It just produces the output. And the output, increasingly, is indistinguishable.
Who’s Actually Using It (Everyone, Unofficially)
The official position of most agencies and creative teams is nuanced: “We use AI as a tool to enhance human creativity, not replace it.” The unofficial position—visible in how projects are actually staffed, what junior copywriters are being hired for, and how much time senior creatives are billing versus claiming—is somewhat different.
The copywriter who used to spend three days developing five routes is now spending one day. The difference is going somewhere—into faster turnaround, into thinner margins, into not hiring the next junior who would have learned on exactly those three days. The productivity gain is real. The redistribution of that gain is almost entirely to the client or the agency’s bottom line, and almost none of it is going to the remaining creative staff who are now expected to produce more with the same headcount.
This is not a conspiracy. It’s just capitalism operating at normal speed. The tools reduce the cost of production. The cost reduction gets extracted before anyone has a chance to negotiate. And the people who relied on the billable hours those tasks generated find themselves in conversations about “evolving their role” and “moving up the value chain”—which sounds like opportunity and often means redundancy with extra steps.
If you’ve been following what’s been happening to junior creatives, none of this is new. The entry-level pipeline is narrowing in real time, and the industry is watching it happen with a combination of market pragmatism and collective guilt that manifests mostly as LinkedIn posts about “the future of creativity.”
The Copy That Nobody Can Tell Was Written by a Machine
Here’s the part the industry doesn’t want to examine too closely: most of the copy that has already been generated by AI and published in the world was not caught. Not by readers. Not by clients. Not by brand teams. The detection tools are unreliable, the tells are diminishing with each model generation, and the humans reviewing the copy are often looking for whether it’s on-brand and error-free—not whether a person typed it.
What does that mean? It means the conversation about AI and creativity has been happening at the wrong level. We’ve been arguing about whether AI can be creative—whether it can produce something surprising, meaningful, original—when the more urgent question is whether the work most brands actually publish requires those qualities in the first place. For a significant portion of it, the honest answer is no. And if the answer is no, the human copywriter producing that work was not being paid for creativity. They were being paid for reliability, speed, and sign-off compatibility. Machines now have all three.
The work that genuinely requires a human—the campaign that has to be culturally precise, the brand voice that lives in a register no prompt has quite figured out, the line that is funny because someone understood exactly the right degree of irony for this specific audience at this specific moment—that work still exists. There’s just less of it than the industry previously required, and it commands less premium than it used to because the floors have dropped.
The Authenticity Problem (Which Was Already a Problem)
There is an irony embedded in this moment that deserves naming. The industry has spent the last decade making “authenticity” one of its primary values—in brand voice, in content strategy, in influencer marketing, in corporate communications. Authentic. Human. Real.
And now the copy that expresses those values is increasingly written by a model trained on the aggregate output of human language, optimised for plausibility and brand consistency. Authenticity was always a performance, of course—always a construct that brands built and maintained. But the current situation makes the performance more visible, which is presumably why nobody is discussing it in the keynotes.
The consumer, for what it’s worth, doesn’t seem to care very much. Engagement metrics on AI-assisted content are not noticeably worse than on human-written content for most categories. The emotional connection people form with brands is apparently not contingent on whether a person typed the headline. Which is either reassuring (the connection was real even if the production method wasn’t) or depressing (the connection was never really there in the first place, just a function of consistency and repetition).
What Actually Changes, and What Doesn’t
The copy will keep getting written. The briefs will keep arriving. The approvals will keep happening. Most of what changes is where the value sits and who captures it.
The creatives who thrive in this environment are not the ones who pretend it isn’t happening and not the ones who hand everything to the machine and call it done. They’re the ones who understand what the machine is genuinely good at—production, variation, speed, plausibility—and what it doesn’t yet have: taste, judgment, the ability to know when a brief is wrong, the instinct to recommend something the client hasn’t asked for because it’s the thing they actually need.
That last part is worth expanding. The most valuable thing a creative can do right now is not write faster or prompt better. It’s to be the person in the room who notices that the brief is solving the wrong problem, or that the tone the client has asked for is going to land wrong with the audience they’re trying to reach, or that the campaign concept is strategically sound but culturally deaf. The machine will not tell you that. It will write you twelve variants of the thing you asked for, none of which will flag that you asked for the wrong thing.
Strategy, judgment, and the courage to push back are not automatable. Knowing when to fuck the brief still requires a human who understands why the brief exists and what it’s missing. That’s the job now. Not the typing.
The Honest Conversation Nobody’s Having
The industry needs to have an honest conversation about what copywriting was and what it’s becoming—not to mourn it, but to price it correctly, train for it accurately, and stop selling clients on craft that has been quietly outsourced to a model.
The alternative is to keep performing the ritual—the research, the routes, the presentation, the revisions—while the actual production happens at a fraction of the declared cost, and the difference is extracted quietly until there’s no one left who knows how to do it any other way. That’s not a future anyone in the industry should be comfortable with.
In the meantime: if you’re a creative building your toolkit for this moment, the Spreadsheet Sloth exists for exactly this situation—tracking what the work actually costs, what you’re actually delivering, and making sure the efficiency gains from your tools end up in your pocket rather than disappearing into a client’s reduced budget expectation. The machine writes the copy. You keep the margin. At least one of you should.
