-
The AI Content Creation Strategy That Has Helped Us Successfully Rank
I have seen quite a few posts going over AI for SEO and content creation. There are mixed feelings. Some people here seem to hate it and others seem to love it. But one thing is for sure, AI is here to stay and you either hop on board or get left behind.
Well, I spent the last few months not only studying prompt engineering thoroughly but going over the best ways to use AI effectively and uniquely for SEO with my team and I believe we have it figured out.
Contrary to popular belief, **1-click content creation or content creation that takes 5 minutes** will likely lead you nowhere. I have not met one person who successfully used it. If you can provide such evidence, I would love to see it.
AI is good, but 1-click content creation will provide you with very similar content to your competitors. At the end of the day, AI is only as good as the person handling it. Trying to streamline the process to the point where there is hardly human input, will very unlikely lead to long term success.
The process takes me roughly 2 to 4 hours for a 2500-word blog post that is fact-checked, SEO-optimised, with internal and external links and every relevant On-page factor.
Assuming you have done your keyword research, organised your clusters and so on, the steps I use are as follows:
**Step 1- Outlining:**
In this step, I hardly use AI, but this is the step I use that allows me to be different in comparison to my competitors. From the top of my head, I create a mind map in a tool such as X-mind that allows me to structure the content on a topic to subtopic basis. Very organised. Once I have literally emptied my mind into it with everything I see as relevant, I proceed to check my competitors’ content structure. I add subheadings they may have that did not occur to me where I find it to be most relevant in terms of order and placement.
I check forums such as Reddit to see whether questions on the topic have been asked and how much traction they got. If so, I format the question into another subheading, I check Amazon for books and look at the table of contents to see the topics they go over. I then proceed to collect the relevant subtopics.
I use ChatGPT as well for more inspiration. I would say that the outline takes me a solid 45 minutes to 1 hour as that is what separates you from everyone else. Your depth, structure, uniqueness, authority and backlinks.
**Step 2- Content Creation**
I structure the content by adding the relevant <h tags> in each topic and subtopic and input the content creation prompt I use on ChatGPT. The content ChatGPT provides is average at best. However, it helps me to have a baseline to expand on.
Once I have a very standard blog post with the contents filled by ChatGPT, I use a tool such as Writesonic to expand on the text in each relevant section and then rewrite it with the tool.
I do this throughout the blog post until the entire blog is concise, informational and unique.
After that, I take each paragraph and run it through 3 different AI-detection tools and 2 plagiarism-detection tools. No less.
The entire process takes me 2 to 3 (sometimes 4) hours from beginning to end for 3000 word blog posts and about half the time for a 1500 word blog post.
As someone who has done copywriting and content editing, I can see the difference in how streamlined our content production has become. It still takes time and effort but 1/3 of the time.
**Conclusion**
Yes, AI content 100% ranks, but it requires human input and human action. A mixture of professional outlining, prompt engineering, ChatGPT or other tools, fact-checking, plagiarism checking, AI content checking and everything else will prove worth it.
-
1 Reply
-
Super interesting post. Thanks. Do you mind telling me what AI-detection tools and plagiarism-detection tools you are using?
-
your post is interesting but my question Is if I use write-sonic or another website for content generating so are the website keywords ranking, good effects on the website if I am using ai generated content in my website?
-
Thank you for sharing. In my personal experience AI content will rank anyway. I have a few blog experiments with 1000+ posts made with AI on a brand new domain. First month 500+ clicks, second about a thousand and it’s growing.
On another, older site with tons of traffic, I spent one month publishing 4 pieces of content per day 100% AI alongside the normal posts we have there. They all ranked all the same, except we had more AI content doing even better on Google Discover
(Including a messed up one in which it invented a movie that doesn’t exist).All of that is using the method: create outline based on a keyword, then generate each section separately. I used N8N to automate. There are a few YouTubers sharing similar experiments with similar results.
I’d say it depends on the niche and, even more, on the target keywords. But it does rank well and, as long as you do the same process of optimisation and link building you’d do for a human written content, I’ll probably be alright 🙂
Edit: typo
-
I guess if you don’t know much about the topic you’re writing about, this approach could work. For me, I can just write an article that is 3,000 long in less than 2 hours anyway, so I don’t really see much point in using LLM 🙂
-
I used to be a content writer and ranked really high, some articles I’ve written last year are still #1. Never used AI, I think this is pretty much some of my strategies.
The only thing I was missing was SEO and keyword research. I didn’t learn keyword research until 2 months ago.
I use AI now, because I work for a digital marketing agency and have many clients. But the writing is complete shit, than how I used to write. but I’ve gotten dependent on it.
Can you share what you do for keyword research or what you do with keywords?
-
Great post. I’d loosely considered doing something similar, though at present my colleagues and I decided against it for the client for which we were considering it.
I’d legit been wondering if what you described was an option. Like you, I create very in-depth, structured outlines for writers. I’d wondered if it was possible to feed something like that into ChatGPT for content creation, and whether doing so — versus just having Jarvis or w/e spit something out entirely on its own — would result in better content quality, while saving time and money.
-
I am still not convinced by AI for blog writing and I’ve spent many hours trying to get it to work with all sorts of different inputs and prompts – not just a typical one liner.
My personal opinion is if you want to use it to brainstorm or produce something fairly generic, it’s fine. At the moment, I am using it to enhance blog posts written by humans, not write them from scratch, because whenever I try to get it to write something fairly original, particularly in fields that I am very well read and familiar with, it fails miserably. The difference between a human writer who is an expert or at least very knowledgeable of the the field they’re writing in and AI at the moment is still quite vast. It doesn’t matter how much I refine the prompt, it just doesn’t compare. Having said that, there are occasions where the AI often writes something that I haven’t thought of already, so it has its use. It will be interesting to see how much it improves in future.
-
So do you rewrite it when you run it through AI detection tools or do you just have the AI tool rewrite it?
What do you think of writesonic though? I’d be interested in checking it out
Edit: Ayo I was just asking a question, who’s out here downvoting ????
-
I would like to build on my own data so it doesn’t hallucinate anything anyone up for helping me out?
-
What A.I detectors shows on your content?
Full A.I generated?And what’s the results you exactly getting from these articles(in numbers) and how old your site is?
Log in to reply.