Businesses aren't buying writing. They're buying an outcome.

I saw a post asking why job descriptions increasingly expect writers to use AI for drafts or support. The argument ran roughly: we’ve been producing great advertising and content without AI for decades, so why start interfering with human creativity now?

Because the economics changed. That’s the uncomfortable answer, and I don’t think there’s a kinder one.

The artistic process isn’t the thing being funded

A novelist can decide that writing every sentence themselves is part of the work. A songwriter can spend six months finding exactly the right line. The process itself carries artistic meaning, and nobody sensible argues otherwise.

Commercial content has a different job.

A business has something it needs to communicate, along with an audience, an objective, a deadline and a budget. It isn’t really buying writing. It’s buying an outcome: understanding, trust, attention, demand, differentiation, search visibility, a clearer explanation of why somebody should care.

Writing is one route to those things. It has been the dominant route for a long time, which is easy to mistake for it being the only one.

So if an experienced content person can use AI to interrogate research, explore arguments, find gaps, structure an idea and produce a good first draft in two hours rather than two days, businesses are going to find that very difficult to ignore. Not because they hate creativity. Because they optimise.

Even in copywriting, the words weren’t the product

The strongest version of the objection comes from short form. A headline. A tagline. Twelve words of body copy that took a week. Surely that’s craft, and surely the hours are the point.

But ask what was being bought. The twelve words, or the work it took to arrive at them? The research, the positioning, the twenty rejected routes, knowing the market well enough to recognise which line would land?

I’d argue it was always the second, and that was true long before AI existed. The words were never the whole value. AI hasn’t changed that. It has just made it much harder to keep pretending otherwise.

I spent years agency side, which is where my B2B grounding came from, so this isn’t a guess. The complicated part was rarely the writing. It was the thinking, and then getting the thinking signed off. Stakeholder rounds. Legal review. The client who prefers their fourth-favourite option. The argument for why the unglamorous line is the right one. None of that is typing.

Which leads somewhere less comfortable. If you’re genuinely better than AI at this, and plenty of people are, that now has to be demonstrated rather than assumed. You’re working inside a business, and the business can see what both options cost. “Mine is better” is a claim, and claims inside organisations get evidenced.

That sounds harsh. I think it’s also the strongest position available, because a demonstrated difference is worth considerably more than a professional assumption.

Identity is not the same as value

I think some of the anxiety here comes from confusing professional identity with commercial value.

For a long time a writer could reasonably say: I’m valuable because I write. I’m increasingly unconvinced that’s where the durable value sits, and I say that as somebody whose entire career came out of writing.

The harder questions are upstream. What is actually worth saying? What evidence supports it? What does this organisation know that nobody else does? Which argument is weak? Which stakeholder request should be resisted? What should this research change? How should the idea move through an article, a video, a sales conversation, a search result? What did the response teach us?

AI has made producing prose dramatically cheaper. It has not made any of those decisions easier. That gap is where the editorial jobs are going, and it’s why the role is shifting from writer towards something closer to editor, strategist, or what I’ve been calling an editorial intelligence operator.

The opposite policy is also stupid

There’s an important caveat, and it cuts against the people who’d otherwise agree with me.

“Every writer must use AI for every first draft” is a bad policy too.

Sometimes writing manually is the fastest way to think. Sometimes AI adds friction rather than removing it. Sometimes its draft is worse than a blank page, and you spend longer arguing with it than you would have spent writing.

The sensible requirement was never use AI. It’s: use the best available combination of human judgement and technology to produce the strongest work at a commercially sensible speed. Today that increasingly includes AI. Tomorrow the tools change again, and the requirement doesn’t.

Defending the right part of the job

Writers should be careful about which part of the work they defend.

Craft matters. Originality matters. Taste matters. Language matters. I’d argue they matter more now that producing competent prose has stopped being scarce, which is the whole case against slop.

But in a business, none of them exist independently of the outcome. They’re how the outcome gets reached, and they’re worth defending on those terms rather than as ends in themselves.

I know this won’t be popular with everybody who writes for a living. I considered softening it and decided not to. It’s where the work is going, and hiding that on a site whose whole argument is about facing evidence honestly would be a strange choice.

Commercial writing can absolutely contain artistry. The artistry just isn’t why the business is paying for it.