Every conversation about AI arrives, sooner or later, at the same question: whose job does this take? It's a fair question. In plenty of industries it has a real and uncomfortable answer, and the people asking it are right to ask.
But the question carries an assumption that's worth making visible: it assumes someone is sitting in the seat. At most early-stage companies, the marketing seat is empty. Nobody is being replaced, because nobody is doing the job.
That sounds like a technicality. It isn't. It changes what the tool has to be good at, and what you should honestly measure it against.
The job exists. Nobody holds it.
Walk into a company of four people and look for the marketer. There isn't one. There's a founder who does marketing the way you do the dishes — late, resentfully, and only once the pile becomes impossible to ignore.
You can spot the empty seat without being told. The website describes a product two pivots ago. The LinkedIn page's most recent post announces something that shipped in spring. The newsletter that was going to go out monthly went out twice, and the second one apologised for the gap.
None of that is a person underperforming. There is no person. The work simply isn't happening — and it isn't happening quietly, which is the dangerous part. Nothing breaks when you don't post. No customer emails to say they noticed your silence. Marketing is the rare function where doing nothing at all raises no alarm; it just produces a slow flattening that's easy to attribute to the market, the quarter, or the season.
“AI takes jobs” doesn't survive contact with an empty seat
Let me be straight about where the fear is legitimate. If you employ someone to write your marketing, a tool that writes marketing is a conversation about that person's work, and it deserves every bit of the seriousness people bring to it. I'm not going to wave that away to sell something. That's a real conversation, and it isn't this one.
What I'm pointing at is narrower. The replacement frame is a subtraction frame: it assumes a fixed quantity of work currently done by someone, and asks who does it afterwards. That arithmetic doesn't run when the current quantity is zero. You cannot subtract from an empty seat. In a company where the seat has never been filled, anything the system produces is net new — work that wasn't happening last month and is happening now.
An empty seat isn't a neutral state
Here's where the comfortable reading goes wrong. “Nobody to replace” sounds like good news — no harm done, nothing lost. It's the opposite. The empty seat has been costing you the entire time. It just never sent an invoice.
Every week it sits empty, the person who would have found you doesn't. The customer who'd have been reminded you exist isn't. The position you'd have built, one post at a time, doesn't get built. Founders discount this because the cost is counterfactual, and counterfactuals don't show up in a dashboard. You never meet the pipeline you didn't build. You just conclude, eventually, that marketing doesn't really work for a company like yours.
What fills an empty seat isn't a replacement. It's a first hire.
This is the part worth taking away, because it changes the bar.
A replacement has to clear a hard and fair bar: better than the professional currently doing it. That's the bar the anxious commentary applies to every AI marketing tool — is it as good as a real marketer? — and it's the right bar when there's a real marketer to compare against.
A first hire clears a different one: better than nothing. Better than the Sunday evening you resent spending on it. Better than the fourth consecutive month of silence.
That is not a licence for bad output, and I'd be suspicious of anyone who used it that way — “better than nothing” is how you end up publishing the generic slop that makes people distrust this whole category. The drafts still have to be good enough that you'd put your name on them, because you're going to. But it does mean the honest comparison for a company with an empty seat isn't Scout versus an agency. It's Scout versus the marketing that is currently, actually happening. Which is none.
Scout isn't trying to out-write the best marketer you could hire if you could afford one and had time to manage them. It's trying to make sure the seat isn't empty on Tuesday.
The seat doesn't stay a seat
The interesting thing about a first hire is that they don't stay new. Neither does this.
It starts by fitting the shape of the role by default — it knows your positioning, your audience, and how you sound before it writes anything, so week one isn't a blank page. Then you start correcting it, and the corrections stick. The edit you make in July is why the September draft doesn't need it. You're not maintaining a tool; you're training a colleague who happens to never forget what you told it.
That's the arc we build against: from an empty seat to part of the team. Not because the software is magic, but because a seat someone occupies every week compounds, and an empty one doesn't.
What this doesn't claim
Two honest limits, because the category has earned its scepticism.
This isn't an argument that AI displaces nobody, anywhere. It's an argument about a specific and very large set of companies — the ones where marketing was never a job anyone held. Where the seat is occupied, ask the harder question, and ask it properly.
And filling the seat doesn't mean the work leaves your plate entirely. Someone still has to decide what's true, what's on-brand, and what ships. That's you, and it should be — a colleague you never review isn't a colleague, it's a liability. What changes is the size of the ask: minutes of judgment instead of hours of starting from nothing.
So the next time the replacement question comes around, notice who's asking and about which company. For a business with a marketing team, it's a real question with real stakes. For the four-person company whose last post was in spring, the answer is nobody — and the better question is how the seat stayed empty long enough that everyone stopped noticing the work wasn't getting done.