Affiliate Disclaimer: This article may contain affiliate links, if you click a link and make a purchase I may receive a commission at no extra cost to you. Learn more.
When you sign up for my newsletter, I send you a welcome email — and I get lots of responses, which is fantastic.
But I’ve noticed something… odd.
I guess it is a sign of the times, but an alarming number of you are using AI content in a risky way.
You send me your sites, and I take a look, and within seconds I can tell that the content is generated with ChatGPT, usually 3.5 (because it’s free).
If I can tell so easily, don’t you think Google can? And is that content actually helpful?
In today’s email, I’m going to show you how you could write better and more helpful content using an AI writing assistant intended to get the best out of you in a more efficient way.
It’s called Reword and is this email’s sponsor:
Reword helps you to write articles that rank with a smart AI assistant.
Fresh content ideas, automatic topic suggestions, facts and statistics all handed to you.
You can even integrate into your WordPress blog so the AI can learn to write more like you.
Get a free 14-day trial by giving it a spin and see for yourself how AI can be used in the right way.
Sign-up 👉 Reword 14-day free trial
So what is it?
Reword is a powerful tool that can help writers create high-quality content that is both engaging and informative.
It’s designed to help you write content that ranks with on-the-fly SEO analysis as well as constant topic suggestions, identifying matching search intent, and more.
Using AI intelligently and ethically, writers can improve the reader experience and grow their website without resorting to spammy tactics.
It’s a writing editor that helps you maximize your output and ideas by helping you come up with ideas or write entire answers to a query.
I decided to give it a real test.
So I’ll pluck out a keyword and attempt to write an article with it.
I’ve chosen the dog niche, as I usually do for public examples, and settled on the keyword: “why does my dog keep licking his paws“.
I’m going to write the article using Reword and see how it performs. Let’s crack on!
I’ve started a new article inside Reword, and entered the keyword as my title exactly, plus a brief objective for the article:
That’s because I know that this is a keyword with volume already, so having an exact match title makes most sense.
Plus, it’s already a long keyword/title, so it’s already approaching the safe character limit for blog article titles.
I start out with a blank slate inside the tool, but I’m immediately offered questions I can ask the AI to help me know what I need to write in this article:
At first, I do want to decide what topics need to be in this article, and that means what structure it needs to have. Ergo; subheadings.
I get some great suggestions that could really help me beef up the relevance of this article:
What is especially evident is that these aren’t some nonsense subheadings I often get from ChatGPT.
When I use my human brain, that can understand and appreciate the concerns of dog owners, these ideas make complete sense.
I start selecting which subheading ideas I think are most relevant, clicking on them in the order in which I want them to appear.
When I write articles, I want to give the most immediately asked questions first and the more obscure later. For example, I’m NOT going for “The Basics of Paw Licking In Dogs” first because it’s potentially too much explaining before giving the reader an actual answer — which means they can get annoyed, and bounce.
As it’s what you might call a “response post”, I’m looking to select around six subheading ideas that will likely easily get me to a 1000-1500 word article.
There’s no exact requirement for word counts, there never is, but I’ve enough subtopics to cover here to make it a good reason to do it.
I’ve selected the subheading ideas in an order that I think most makes sense for a dog owner who might lead from one question or thought process to the next:
Why does my dog lick his paws -> Is it a bad thing? -> How do I prevent it? -> Prevention methods -> Seek professional help
It makes most sense to me in this order.
So now I have my structure for a brand new article, and it took literally less than a minute.
Now let’s actually write it.
The way I’d like to do this is by getting involved in the introduction of the article myself and define a good starting point.
Then from there, hopefully the AI can take that information and write the rest.
AI writers often get a bit lost inside long winding introductions and I want to keep it neat and tidy to a method which I know works. That method is as simple as identifying the readers problem, agitating it, and then offering a solution, nice and quickly:
I’ve started my snappy introduction and then opened the Command feature inside Reword to ask the AI to write the “featured snippet” answer block.
I want the readers to fall onto the article and get some kind of ready solution upfront.
The exact phrase I entered was:
“Add one concise paragraph that directly answers the question “why does my dog keep licking his paws?” in 50 words or less.”
I was pleasantly surprised by the output:
Now I’ll move onto each subheading and get the AI to write the rest for me.
All I need to do is select the subheading and then use the Enrich feature of Reword:
And as if by magic, it’s off to the races, writing it for me:
Granted, there’s not a ton of information for each.
But I can easily improve that by selecting these sections, and asking it to Enrich again.
You also have to keep in mind that perhaps simpler is better.
Not every reader needs a history lesson about dog parasites and the thousands of types of them. They just want to know what may be some of the causes, so they can understand what they’re dealing with.
As I moved on to the next subheading, I discovered that the AI can’t always get it right:
What I was expecting was for the AI to answer this query succinctly in this introductory paragraph.
But instead, it waffled on a bit and then changed the topic to something else.
Let’s be real, AI is never going to perfect and this is one of the reasons why you need to guide it and tell it what you want, sometimes explicitly.
This is why you shouldn’t expect to get great articles out of one-click AI writers!!
What I want is a direct answer to this subheading question, then we’ll elaborate.
So I’ll just ask it for that, this time in a Command:
The full prompt I use is:
“Directly answer the question “is your dog licking its paws too much?” and then elaborate on what might be considered “too much” of dog licking its paws.”
My aim with this prompt is to get an immediate answer to the subheading, which has some usefulness for snippet capture, but mostly it’s to just make the content incredibly helpful.
Here’s what I got:
Wow, now we’re talking!
As I continue through creating the article, I find that using Enrich from a subheading doesn’t seem to be the best route, because I end up with several sections of content that I don’t want.
For example, it’ll try to add a final subheading every time about seeking professional help — something that I want the article to do later, at the end.
So, instead, it’s best to get your own idea of what output you want to see and use the Command to get it. Which it’s really good at!
But there’s more features in this tool.
Like I can use the Simplify feature to bring down over-explained section introductions:
Which gives something a bit better:
And while I’m reading the article, maybe there’s just something I don’t like the way it’s phrased.
Maybe it feels too robotic, or too much like lame copywriting that “over does it”, I can use the Rephrase feature:
These are all helpful features to increase your content output and improve it efficiently.
But there are more powerful features inside Reword.
In particular, the Ask any question feature has a prefilled option that is always a great addition to an article to validate improved research:
And this presents me with snippet statistics, plus their sources:
I just scan my article and look for spots where one of these statistics might be a relevant inclusion, and click to add:
Now I have an article that’s complete and ready to publish.
If I want, I can invite teammates — like editors or formatters — into the document to run their own checks.
And before I publish, I can check the Performance tab to find out how well optimized this article is for search:
There’s definitely still some work to do to make it fully SEO optimized, but I’m happy that all the reader intents have been met in the content.
For me, matching the reader/search intent is the most important thing to winning position on the SERP but also pleasing users that do find your content.
One final check to do that some of you might find important (the jury is still out on this for me), which is to scan the content in an AI & plagiarism checker:
This is fairly accurate to how I wrote this article so far, which was mostly manual on the introduction and generated with everything else.
But does this even matter?
There are many perspectives on whether AI checkers are of any use at all, similar to the debate on whether Google likes or dislikes AI generated content, considering Google is “OK” with AI content as long as it’s helpful.
What we know is always true, content must be helpful. And if it does that, it’s doing what it’s supposed to do.
But one thing I think is often true with AI generated content, is that in time it can become less helpful than other original content on the SERP which can lead to bounces off your page, and that is something Google grades highly in ranking.
Anyway, the article is ready to publish!
I’ve published the full article on my blog for you to take a look.
What do you think? Is this content helpful, useful, and would it rank?
The blog post obviously could benefit from a lot more effort put into it, but it’s a simple example for the purpose of this tutorial.
I was able to complete this article in about 15 minutes (when I disregard time spent writing this email).
I think that having topic suggestions inside your editor is a really powerful feature that enables you to start writing straight away instead of having to go to the SERP for the keyword and reading tons of other articles just to get up to speed yourself.
The ability to generate content when your own brain is stumped is really good, and the output is decent, but you should probably always apply your own manual edits.
The key pitfall of using AI is when you get lazy with it.
If you use AI in an intelligent way, like as an actual assistant rather than a pure source, then you’re more likely to produce better content that can continue to rank.
It’s still early for me in using Reword and I continue to be hesitant using AI for actual content on my main site.
I do use ChatGPT 4 for certain tasks like compiling information or summarizing product details, but I am careful about easing into it.
As you should be.
But I do think, and have said multiple times previously, the way you use AI is everything.
To those of you that are spitting out ChatGPT 3.5 articles after giving it an outline, or using a one-click AI writer, I don’t think you should expect long-term results.
AI can be a friend or foe, it’s all about how you use it.
I even set up a test niche site using ChatGPT 3.5 with 70+ articles in a cluster:
I am already seeing ranking results, likely due to the tight topical clustering and interlinking created on it.
But do I expect it to work mid-term, or long-term? Nope.
What I could do, though, is use something like Reword to come back to that project after 90 days to improve every article with a little more human guidance.
I think Reword can be a powerful assistant in either rewriting AI generated content to add more value and helpfulness, or a great tool for improving pre-existing content.
If you want to give it a whirl yourself, you can sign-up for a free 14-day trial here.