The robots aren't coming. They've already moved in

I watched a junior marketer on the Luas, building a campaign in the notes app. He had one earbud in, the other dangling, and he was flicking between a prompt, a doc, and a half-finished landing page like it was nothing. At one point the train jolted, his thumb hit send, and the tool still spat back three usable headlines before we got to the next stop. It was weirdly calm. Not heroic. Just normal.

That’s what messes with your head. The work isn’t brilliant, but it is fast enough to change who needs to be in the room.

For years we built marketing like a relay race. Strategy hands to creative. Creative hands to design. Design hands to web. Web hands to analytics. Someone somewhere produces a deck to prove it’s all connected. Each handoff adds cost, time, and a new chance to dilute the point.

Now a lot of the relay baton work is done by software. Drafts, variations, basic layouts, summaries, first-pass analysis, the endless “make it shorter, make it punchier, now do ten versions.” That used to be paid labour. Now it’s a button. So the threat isn’t that “robots will take your job” in some dramatic overnight way. The threat is that the middle of the org chart, the part built to move work along and tidy it up, gets thinner because there’s less to move and less to tidy.

You can feel it already in the kind of tasks that used to justify whole roles. First draft copy. Basic competitor tables. Social calendars. Slide clean-up. Even the polite meeting notes that everyone pretends to read. They can still be done by people, but the economic argument for needing so many people to do them has cracked. Most teams haven’t faced the real implication yet.

They’re still improving workflows that exist mainly because execution used to be slow, expensive, and annoying. They’re still hiring for narrow slices of output, even as output becomes the cheapest thing in the room.

This doesn’t mean the work goes away. It means the shape of the work changes. A smaller team can now ship a first version without waiting two weeks for the slot, the resource, the review, the rewrite, the reschedule. That speed is not automatically good, but it is real. And it forces a question most marketing teams don’t like answering, which is: if the work can be produced quickly, what exactly are we being paid for. The answer isn’t “more content.” Nobody is crying out for more content. The answer is direction. Taste. The ability to make a call under uncertainty and stand over it. The ability to say no to the shiny wrong thing. The ability to spot the quiet risk before it becomes a public one.

Software will happily produce a hundred plausible routes. It will not carry the consequences of choosing the wrong one. It won’t sit in the room when sales are down and the board asks why. It won’t be the person whose name is on the decision.

That’s why teams that win won’t just “use AI.” Everyone is doing that now. The teams that win will rebuild around fewer handoffs and clearer ownership. Fewer people passing work along, more people capable of taking a piece end to end, from question to draft to test to fix.

The robots have already moved in.