Why AI fails to create audience-centric content

Claire Heginbotham
September 13, 2023
3 mins
Share this post

When I introduce myself to people as a writer, I know that in 30 seconds, the "sooooo, what about AI?" question is coming my way.

A rather roundabout way to avoid saying what's truly on their mind: Is AI going to wipe out writing as a job?

Hello apocalypse.

The first few times I got asked about AI, I was confused about how to answer. In the next few months, I grew almost offended. But now, nearly a full year after ChatGPT was released to the public —  I tilt back my head and confidently laugh.

Because I would love it if ChatGPT could do my job.

Writing is incredibly hard. And I pride myself on being strategically lazy. If I could outsource one of my biggest challenges to a robot, I would.

Why do we feel that an AI response = content writing?

There are many reasons why people think that AI is dangerous for the writing industry, but to my mind, there's one very human reason that takes the cake:

Everyone, including you, knows they can write.

I can write well enough, you think; after all, I've written countless essays in school and thousands of emails to important people. It takes a tiny leap of faith for one to ponder, if everyone in the world is a writer, how hard could it be for AI to be a writer too?

It is those teensy suspicions whispering at night that slowly take hold and morph into the nightmare that you, your job, and your way of life could be overturned in a matter of months.

Fear not dear reader. Because while everyone can write sentences, very few can write coherent paragraphs, and even fewer can make a convincing written argument.

And the same goes for AI. Just because it writes, doesn't mean it writes well.

Earlier, I said writing is hard. And that's because — along with basic skill — it requires you to combine knowledge about specific audience segments by using two opposing abilities; creativity and logic. And until AI can do either, I think my job is very, very safe. A recent study conducted by SparkToro confirms that people are quickly realizing AI is no magic trick and regular use has been steadily falling since April 2023.

Now you know the limitations, here's what AI can actually help you with

All writing that sells something — whether it sells entertainment or education — needs to convince the right people that it's worth reading. And AI really struggles to understand buyer needs like you and I. Instead of a writer, think of AI as a writing tool. Something to assist and automate, but not to DO.

Any writer worth their salt will likely create in three stages. You can use AI in each of these stages for everything from blog posts to landing pages to sales emails.

At our agency, we've named the stages of content creation:

  • Outlining: Where you flesh out what you want to say in headings and bullet points.
  • Writing: When you create a first draft of the actual written product.
  • Editing: Where you edit the first, second, and even third drafts into something readable.

How to use AI for outlining

Basic keyword research and audience insights can be uncovered by asking ChatGPT about the subject. It's preliminary research into what's currently out there and what preconceived buyer personas might look like.

We tell it about the audience segments, then ask it to outline posts or paragraphs, usually ignoring the structure it provides, but checking the information it considers as 'important' for the topic. Then, we do some manual research and write a real outline.

If I've got subject matter expertise, then I find using AI for outlining actually hinders the process. Check out how boring this post would have been if I used AI to write the outline.


This is subjective, so how much you lean on AI depends on you. For me, I use it when I want to describe something simple fast. Basically, simple one-sentence answers on well-documented topics. I usually edit almost all of these sentences to flow better with my actual content. Here's what this looked like recently.

Sometimes I want a bit more. Recently I used Ahref's voice and tone technique to get AI to write me some emails inviting people to fill out a survey. If you read through this chat, you'll see what it outputted, and how I STILL changed a fair amount of what was written to make a stronger argument for our target audience.

Or I want it to simplify something that we've already written. Here's another example.

Again, I ended up editing almost every single line there. As it is, it reads like something written by a total amateur.

Is using AI faster than just writing it from scratch? The jury's still out on that one ;)


The only true AI that makes it into our editing process is Grammarly, used for spelling, grammar, and tone.

I also use keyword optimizers that say they use AI to identify key phrases, but it's a pretty mechanical process, not something I'd call AI — it's more automation than anything.

Otherwise, AI can't actually do what I'm looking for in editing. I'm using my experience to sweep for clarity, specificity, logic, and so much more that it could be another post.

How to take the audience-centric approach with your content marketing

Text-based search engines are out and social media sharing is in. You need to create content that your audience needs (and wants) to read, share, and talk about. If you want to understand more about how this all works, Amanda Natividad over at SparkToro has written a fantastic guide on how audience-centric marketing is your new growth flywheel.

In a nutshell, when doing content marketing with the goal of maximum shareability and site traffic, you need to empathize with your reader (or target audience, viewers, users, whatever you prefer).

Let's say I was writing a post for you specifically. This is the kind of audience insights that I'd want to gather/understand about you before I even open a word doc:

  • How deeply you understand the topic of the post
  • The jargon you use regarding that topic
  • Slang you might know and use
  • Your sense of humor and pop-culture knowledge
  • Your preferred reading level  
  • How you've talked about your pain points
  • How you might feel about the topic of the post
  • What your customer journey maps might look like with specific products/features
  • What you might know about the industry or product mentioned in the post
  • What you might have already read if you're coming from Google
  • What preconceptions you have (eg. everyone can write)

Once I have the answers to this list, I have a fairly strong base to write something that you'll resonate with. With all this data to use, I can think. Not just about the topic, but about you, the reader's, relation to the topic.

And that's how I (and many other writers) develop clever analogies, interesting ideas, and captivating stories.

Copy Island creates content that captures, converts, and retains your customers.

Subscribe to our newsletter

Book an intro call / send an email to claire@copyisland.com

Read our latest case study for blog content

Share this post
Claire Heginbotham
Copywriter at Copy Island