I was making some Instagram videos this week and realised something about the way I’ve been talking about AI content.
I think I’ve been using one of my own early problems to explain yours. And they’re not the same.
When I first got properly into AI, I was still figuring out a lot about my business, what I thought and, most importantly, learning new tech skills. I’d give it an idea I was developing for my system, or a lesson I thought was valuable to share, and with one click expect it to turn that into a week’s LinkedIn content.
Most of the time, there just wasn’t enough of my input around the idea yet because I hadn’t fully worked it out myself.
That genuinely was part of my learning curve. But it isn’t a fair place for me to start when I’m talking to an established business owner.
Your learning curve is different to what mine was.
If you’ve been doing your work for years, you’re starting somewhere else. Put you in a conversation with a client and you can explain why you do something a certain way. You know where people get stuck. You know what you’ve seen work and what you would push back on. There are years of experience behind what you’re saying.
That’s why telling you to develop better ideas doesn’t sit right with me anymore. I’ve realised it’s not that you have nothing useful to say.
You already have what AI needs.
So the question I’ve been thinking about is: if that’s true, why can AI still give you content that feels generic? And why can you still end up so involved in correcting it afterwards?
This is where I think “give AI more context” gets a bit vague.
Of course context helps. I want AI to know who I’m talking to, what my business does and how I tend to explain things. But it can know all of that and still not know the important detail behind the thing I want to say today.
Say you want to write about something you keep seeing clients get wrong. AI might understand the broad problem. It still doesn’t know the particular example you’re thinking of, the bit people usually misunderstand, or what you would naturally say if someone asked you about it in a conversation.
Those details have to come from you. They’re the specific details about the content you want to write that AI would never know, and it should never be allowed to invent them.
The problem is, AI can still write without them. It has enough general information and enough language to make something sound perfectly plausible.
When those details are missing, AI just fills it in with average words. And there’s your generic content.
That’s the bit I think matters. It isn’t that the person behind the content has no expertise. Quite frankly, it’s the opposite. The expertise is there, it just hasn’t made it into this particular piece before the writing starts.
Then you end up putting it back in afterwards.
You read the draft and think, that isn’t quite what I meant. There’s an example missing that you would have said naturally. A line sounds like something you would never say. So you correct it, add the detail, pull the wording back and keep going until it feels like yours again.
I know that part well because I’ve done it… many times. I’ve had AI give me something quickly, then found myself swapping words, rewriting lines, struggling to make it just sound anything like me.
What I’ve changed since then is where I want to be involved.
I don’t want to wait until there’s a full draft before I find out what AI has misunderstood or filled in for itself. I’d rather talk the idea through first. I can explain what I mean in my own words, let AI reflect back what it thinks I’m saying, notice what’s missing and add the detail it couldn’t possibly know. Then I can ask it to write.
That does mean there is human work before the draft… sorry to dash your hopes. I know that can sound backwards when the whole reason for using AI was to save time.
But the work doesn’t disappear just because I move straight to “write this for me”. If the important substance behind my content idea isn’t there at the beginning, I often end up doing that work later anyway, except now I’m doing it while trying to rescue something that has already been written around the gaps.
This is what I mean when I talk about moving the effort upstream.
There’s another reason I think that matters for an established business. You don’t create one piece of content and then you’re finished. There’s always another newsletter, another post or follow-up email you need to write because staying visible and useful is part of the work.
If the thinking behind an idea has been captured properly, it doesn’t have to disappear into the first thing you publish. AI now has something dependable to come back to. The next piece still needs its own shaping, but you’re not rebuilding the meaning from zero or asking AI to guess the important bits all over again.
Your thinking travels. Your expertise travels.
Not because the goal is to churn out more content for the sake of it. It’s because the content you already need to create to stay visible and build trust has a better chance of carrying what you actually know.
I think that’s the reframe I had been missing when I looked at my own learning curve and yours as though they were the same.
If you’re an established business owner, it isn’t that you need better ideas before AI can help you. You already have what AI needs.
Your learning curve is about getting the important parts of what you know into the process before AI starts writing, so it has something real to work with instead of making up the bits in between.
