“AI won’t take your job. But someone who knows how to use it… probably will.”

I’ve been nodding along with that since ChatGPT burst onto the scene.

But agreeing is one thing. Actually seeing it is another.

It rang true last month when I was knee deep in marketing strategy. I had a tight deadline and half a dozen stakeholder notes spread across emails and slides. On a different day I’d have chipped away for hours trying to find the thread. Instead, I shaped a prompt for ten minutes and pressed go.

What came back wasn’t perfect. A few clichés and a leap in logic here and there. But it was close enough to give me a great start. Maybe a touch more than seventy per cent there. however, I wasn’t staring at a blank page. I was editing, cutting, and improving. It felt like somebody had done the heavy lifting and left the interesting parts for me.

That’s the point I keep coming back to. AI doesn’t replace thinking, it compresses it. It moves you through the swamp faster so you can spend time on the bits that matter.

And in strategy marketing, where time, clarity and structure matter, that is an edge.

‍

What actually changes when you add AI?

It doesn’t take away your judgement, and it shouldn’t touch your taste. Those are still the things that matter most. What shifts is how the work feels once you bring it in.

For me, the biggest change is in the early stages. I used to spend hours just pulling stuff together, notes from stakeholders, scraps of research, fragments of old campaigns. Now that part only takes minutes. The raw material comes back in a shape I can actually use, which means I get to start thinking earlier.

The flow of ideas changes too. In the past I might have chased one direction hard, because getting alternatives took too much time. Now it’s easy to see three or four routes side by side. Not all of them are strong, but one usually holds something worth keeping.

And then there’s testing, ideas don’t need weeks to be “proven” anymore. You can pressure-test them almost immediately. Ask the tool to play the sceptic, to poke holes, to surface objections you haven’t thought of yet.

None of this is glamorous and sometimes it still feels clumsy and occasionally it makes things messier before they get clearer, but overall the momentum is different. You move forward faster, even if the path is a little uneven.

‍

Where AI already earns its keep in marketing strategy

Here’s what I’ve found so far. It doesn’t replace the work, it just clears the boring parts and gives you more space for the good stuff.

Organising messy inputs
Feed it notes, research highlights, call transcripts. Get back a one-page summary with tensions and gaps. Then push it: what’s missing, and why does that matter?

Audience insights
Rough personas, jobs to be done, buying triggers. Not perfect, not finished. Just scaffolding you can test and refine.

Creative breadth
Ten headline angles. A handful of metaphors or three routes for a campaign story. Most will be likely be average but one will spark something and that’s enough.

Briefs, not outcomes
Probably the best use for me. Turn chaos into a clear brief with background, objective, mandatories, tone, success criteria. Then work on it to make it real.

‍

The uncomfortable bits

Sometimes AI feels like cheating... the first draft is so tidy you worry you are becoming lazy. Other days it hallucinates a confident paragraph that is wrong in three different ways.

That is fine. The point is not to accept, it is to interrogate. Ask for sources. Ask what would change the answer. Ask for the strongest argument against the suggestion. If it cannot justify itself, throw the paragraph away and write your own. Your taste still runs the show.

Also, a mild contradiction I live with: I want originality, yet I ask a model trained on the internet to give me a starting point. The way through is simple enough. Use it for breadth, not final polish. Let it widen the angle, then narrow it with your judgement.

‍

A simple playbook you can steal

This isn't a grand system, it's just a tidy loop I keep coming back to.

1. Frame the job
   One paragraph on goals, constraints and audience. Be plain. If you cannot explain the ask clearly to a model, you probably cannot explain it to a colleague.
2. Load the context
   Bullet the inputs you have: research, notes, metrics, examples to avoid. If you are short on data, say so. Ask what would be useful to gather next.
3. Request options, not answers
   Ask for three routes, not one. Different angles, different trade‑offs. Force diversity with instructions like “one conservative, one bold, one surprising”.
4. Pressure‑test
   For the best route, ask “Where will this fail?” and “What would a sceptical stakeholder say?”. Capture the objections. Improve the thing.
5. Human pass
   Edit as if a junior handed you the draft. Cut the fat. Add the line that makes it sing. Replace generic claims with proof.
6. Decide the next smallest step
   Not the whole rollout. The next test. An email variant, a small ad set, a landing page block. Let feedback shape the rest.

‍

What stays human

Taste. The bit where you know a line will land with a particular audience. The moment you choose not to say the smart thing because it is not the right thing. Negotiating priorities when two stakeholders want the opposite outcome. Handling the awkward conversation when the data says the pet project is not working.

Also craft. The sentence rhythm, the small metaphor, the restraint. I still care about those. Maybe more than before.

‍

Guardrails That Keep Me Honest

First, never outsource ethics. If a use case feels wrong, it probably is.

Second, protect privacy. Strip out sensitive data and be cautious with what you feed into third-party systems.

Third, verify facts. If a claim matters, check it. If a quote matters, source it.

Fourth, document good prompts. A working structure today will save you hours tomorrow.

And finally, default to clarity. Clever but confusing outputs don’t help anyone. Rewrite in plain English.

‍

So, is AI coming for your job?

I don’t think so. Not in the way people fear. The job is simply becoming something new. A little more orchestration, a little less manual stitching. Fewer hours spent turning rough notes into neat paragraphs. More hours deciding what matters, and how to prove it.

If you are a marketer who learns to wield these tools, you will move faster than the version of you who doesn’t. You will probably move faster than your peers too. Quietly at first. Then obviously.

And if the model hands you a nonsense paragraph now and then, good. That is where judgement steps in. That is where your taste earns its keep.

Perhaps that is the real shift. Not that AI is taking the work away, but that it is giving you a better shot at doing the part only you can do.

Small edges compound. That is how people “who know how to use it” quietly pull ahead.

And yes, I am still learning. I suspect we all are.

‍