Last week I told you about the detours. The data centers, the asphalt, the Korean boss, the slow climb back to marketing. This week, I tell you what happened when I stopped reading about AI and started actually poking around in it.
From Curiosity to Questions: How I Started Using AI as a Content Marketer
It started with a rabbit hole (it usually does, for me at least).
I forget the exact article, but the gist was this: companies were using AI to do in hours what used to take entire marketing teams weeks.
Some were cutting headcount. Some were replacing roles that had taken people years to build. The headline was written to make you anxious.
For me, it made me curious instead, which I am not sure says anything good about me.
So I started poking around.
The first tool I tried was ChatGPT. As with most people, I suspect. And my honest assessment after the first few weeks was: fine, but nothing special. I was asking it to write posts, knock out short copy, that sort of thing. The output was serviceable in the way that a meal from a vending machine is serviceable. It filled the gap but you would not write home about it.
What I did not understand then, and only figured out later, was that the problem was not the tool.
It was me.
I had no idea how to prompt properly. I was not giving it context. I was not training it to understand my voice or the audience I was writing for. I was essentially handing a capable new hire a one-line brief and then complaining when the work came back generic.
That is not an AI problem. That is a management problem, and I have made that mistake before with actual humans too.
By the time I found my way to Claude and Gemini, I was a better “boss”. I had done enough reading, made enough mistakes, and learned enough about context and framing to actually get something useful back.
The tools had also improved, which helped, but the bigger variable was that I had stopped treating AI like a search engine and started treating it more like a collaborator with a specific skill set and some notable blind spots.
That shift took longer than I would like to admit.
Somewhere in the middle of all this, I started having ideas. Not just “how do I use this tool better” ideas, but actual build ideas. Things I wanted to exist that did not seem to exist yet, or that existed but cost more than the arm and leg I was willing to do without.
One of them, and I realise this is going to sound like a tangent, was a better system for reading the Bible.
I know how that sounds, but bear with me.
Wait, stop, don’t go! I am not turning this into a faith post.
The point is that I had spent years trying different reading plans, apps, study systems, and none of them quite did what I wanted. They were either too rigid, or too shallow, or built for a different kind of reader than I was. And at some point it occurred to me that with the right setup, AI could help me build something closer to what I actually needed.
Python, which I use at a level I would charitably describe as slightly advanced basic, was about the extent of my programming knowledge. But the thought stuck.
The question of what I could build, rather than what I could buy or subscribe to, was new. It felt different from anything I had asked myself in a long time.
It’s a whole, new world, and it felt like the early internet again. That period when the gap between having an idea and actually making something was getting thinner every month.
Except this time I was older, more beat up, and carrying more than two decades of work experience across industries that had nothing obvious to do with each other.
Which, it turns out, might actually matter.
What Solo Content Marketers Actually Need from AI (and What They Don’t)
Here is the thing about spending years across industries that seem unrelated: at some point you stop seeing separate fields and start seeing the same problems wearing different clothes.
I am not going to dress that up as a gift.
For most of my working life it absolutely felt like the opposite.
But the specific frustrations I want to explore with AI tools do not come from trend reports. They come from real jobs, and from one problem in particular that I ran into more times than I can count.
Content marketing, especially in B2B tech, is a lonely discipline.
I do not mean that in an existential sense. I mean it practically.
You are often the only person in the room whose job is to think about what the company says and how it says it. You are surrounded by people who have opinions about the output but very little time to help you produce it.
Sales teams are always busy. You know you need to understand what questions their clients are actually asking, what objections keep coming up, what language the market is using. That information lives inside the heads of your colleagues. Getting it out requires catching someone at the right moment, which in my experience happens about as often as it should but never when you need it.
So you improvise.
You work with what you have. You write copy based on half a conversation you managed to have last Tuesday and a product brief that was last updated six months ago. You make editorial calls that should be validated by someone closer to the customer, and you make them alone because the alternative is waiting, and waiting means missing the deadline.
Then there is the other problem, which is having no one to think with.
Not someone to approve your work.
Not someone to review the final draft.
Someone to sit with (preferably with a strong cup of coffee) while you are still in the messy middle of an idea and say: does this angle make sense, or am I about to walk the brand into a corner? Is this the right framing for this audience, or does it read differently to someone who actually buys this product?
I used to solve this by walking over to someone’s desk and talking it through. Open plan offices were useful for that, whatever else you might say about them. Remote work and busy calendars have made it harder. And even when you could do it, the person you were talking to was doing you a favor. You could feel them waiting to get back to their own work.
AI handles this particular problem better than I expected.
Not because it knows your market. It does not, not without you teaching it.
But because it is available, it is patient, and it will push back if you ask it to. I can now run three or four different angles on a content idea before I commit to one. I can test a hypothesis against a sceptical audience by asking Claude to argue the other side. I can check whether a piece of editorial logic holds up before I spend two hours writing the full draft.
That is not AI replacing editorial judgment.
That is AI giving the solo content marketer a thinking partner at two in the afternoon when everyone else is in a meeting.
The output is still mine. The decisions are still mine. The brand stays mine to protect or damage. But I am making those calls with more information than I used to have, and fewer moments of staring at a draft wondering if I am the only person who thinks this angle is a good idea. Seven years of doing this work has taught me what questions to ask. AI has given me somewhere to ask them between nine and five when the sales team is unavailable.
That combination, editorial judgment built over time paired with a tool patient enough to work through the thinking with you, is what I am trying to get better at. It is also, I suspect, what separates content marketers who will continue to be useful from the ones who will not.
Why Publishing More AI Content Is Not the Answer
Publishing content that does not land is one of the stranger professional experiences I know.
You write something. You spend real time on it. You believe in the angle, or at least you have talked yourself into believing it enough to hit publish.
And then it goes out and nothing happens. Not a wave of negative feedback. Not engagement.
Just silence.
The way I have always described it: you are standing in a large room full of people, shouting, and not one person turns around. Nobody tells you to be quiet. Nobody argues with you. They just carry on talking to each other as if you are not there.
After a while, that silence does something to you.
Not dramatically. Slowly. You start second-guessing the angles. You wonder if the problem is the writing, or the platform, or the audience, or the product, or some combination of all four. You publish more, hoping volume fixes what quality apparently could not.
It rarely does.
I have been in that room more than once.
I imagine most content marketers have.
What interests me now is whether AI changes any of that.
Not by generating more content faster. The brands doing that are mostly just shouting louder into the same silent room, and some of them are going to figure that out only after real damage has been done.
What interests me is whether the thinking-partner use, the hypothesis testing, the editorial pressure-checking before you commit to an angle, produces content that is more likely to make someone turn around.
I do not know the answer yet.
That is what I am trying to find out.
If you are a brand trying to figure out the same thing, or someone who hires the people who figure it out, this series is probably worth following. Not because I have it solved. Because I have spent enough years on both sides of this problem to ask the right questions about it.
Next week, I’ll talk about the brands publishing AI content at scale right now and quietly wondering why nothing is working the way it used to. I have an idea on what is going wrong. Come and let me know if I am right or wrong.
Isaiah Chua is a B2B content marketer based in Singapore with seven years of experience across cybersecurity, fintech, and enterprise technology. He builds AI-assisted content workflows that help marketing teams publish faster without losing editorial judgment. If you are building a content marketing function and need someone who can think across verticals, write with intent, and understand what content is actually supposed to do for a business, his portfolio is at ikechua.com.