How to Write a (Good!) Blog with AI - for Copywriters
Writing quality blog copy with AI isn’t instant — but it’ll still save you a TON of time.
written by Sarah Check & AI Sarah
The world didn’t need another AI tool, but I did.
In 2021, I was sitting in a Cracker Barrel, pregnant, talking with my husband. My copywriting business was running well – but with a baby on the way, I knew I’d need to speed up my process. By a lot.
I told him, “There’s a lot of copywriting that doesn’t need my original input—like cleaning up voice dictation automatically. That would be amazing if AI could help with that kind of stuff.”
At the time, ChatGPT didn’t exist yet. I tried a few AI writing tools, and while they worked for generic content, they didn’t make things faster when I needed something original or optimized for SEO.
Then, when ChatGPT launched, I tried using it both as-is and through GPTs made for blogging.
It wasn’t much better. I was still heavily editing to remove the “AI sound.” And worse – the blogs flagged plagiarism checkers pretty bad. A major problem for SEO (which is often the whole point of writing blogs).
So, I came up with my own process for using AI in blogging, which I’m going to share with you in this blog.
I know it works because I was able to write eight long-form blogs while nine months pregnant and taking care of a toddler. These 2,000- to 3,000-word SEO blogs would have taken me at least five to eight hours each without AI… but I finished each in an average of two and a half hours. And three months after publishing, they were already meeting my SEO goals.
Here’s what I’ll cover in this blog:
Doing Quick Outlining & Research. It’s easy to start, keeps you organized, and give the AI you use some direction.
Start Writing with AI by Drafting Section by Section. Your blog will be longer, be easier to adjust, and higher quality.
Edit and Finalize Outside AI. The more high-quality direction you give when you’re talking with the AI tool, the less editing you’ll have to do.
I’ll also give you a POV of how I write a blog with AI, step by step. You’ll be able to open up ChatGPT, Claude or whatever AI tool you prefer and start up a smashing blog post. For you or your clients.
Let’s go!
1. Preparing to Blog with AI
How do you write a blog with AI? Well, not by prompting it “write a blog about X.”
Even if you load up the prompt with details, like “viral worthy,” “for x niche” “with unique insights,”... you’re going to get a 400ish word piece of slop. It’ll sound like everyone else that used those prompts they got from bropreneurs on TikTok.
People won’t want to read it. Google also won’t be impressed.
How’s an AI supposed to know what will make something viral in your unique niche? And how original are the insights going to be if it’s just cobbled together from what it found on the internet?
We’re going to write like a professional copywriter. With an outline. It’s going to make things way better quality with very little effort.
Creating an outline kills a bunch of birds with one stone. It gives your blog structure, which makes it easier to write and read. It also helps the AI understand the end product you’re looking for.
Here’s what works for me:
Pull up a Google doc. Write down a simple bullet list of points.
Don’t worry about organizing or having the exact number of points at first. These will eventually act as your core topics. More points will make your blog longer, and vice versa. In my experience, 8-10 points usually results in a 2,000-word blog.
Have thoughts you want included in your blog? Add them to the relevant point.
If you outlined in a doc, download it as a pdf or copy the text. You’ll give it to your AI tool later, and AI will help you refine the outline.
Without an outline, AI often produces short, repetitive content.
If I ask ChatGPT to lengthen a blog, without an outline, it tends to max out at around 800 words. But with an outline, the AI can expand each point into a complete, detailed section.
Don’t skip outlining. Even spending 5 minutes coming up with 8 points dramatically improves the whole process. It’s essential if you want to know how to write a blog with AI that’s both thorough and readable.
Cool.
How else do we prepare to make a blog with AI?
With a little SEO research.
Can AI do SEO research? Yes.
Can it do good SEO research? Depends on who’s at the wheel.
I do NOT recommend seeding your prompts with requests for content that’ll “go viral” “rank #1 on Google” “get lots of traffic” “use top industry keywords.”
Way too vague. That's like a bad boss telling an employee to write something that’ll “go viral.” The request in itself isn’t actionable.
Instead, here’s some of my favorite quick ways to do SEO research for an AI blog:
Have a guess at a long tail keyword (3-5 word phrase) that’s relevant. Google it, then see how it’s worded in the “people also ask” section.
Come up with a relevant long-tail keyword. Look it up on Ubersuggest, and use the variations that have higher traffic.
Look up 1-3 highest ranking blog posts that are similar to the one you’re trying to produce. Feed the blog to an AI tool (as a pdf, url, copy/paste, or web search). Ask it what longtail keywords the blog is using. Use those.
Nice. We’ve got an outline and a longtail keyword (or a few) we want to use.
Let’s start writing this blog.
2. Drafting with AI, Section by Section
Now we’re getting to what you came for: how to WRITE a blog with AI.
That outline you made is basically a section-by-section fill-in-the-blank. WAY easier for you (and AI!) to work with than a blank sheet of paper.
Plus, working section-by-section (instead of all in one go) makes the experience much more manageable. Instead of re-reading an entire blog repeatedly to request edits, I can focus on one section at a time, making specific requests for each chunk.
To start drafting your blog, pick your AI tool. ChatGPT. Perplexity. Claude. Whatever. I’ll shoot straight with you – to put together a top-tier blog, you’re going to need to do a bunch of follow-up prompts, especially with the section-by-section method. So you’ll probably want a paid tool, not a free one with a tight prompt limit.
Here’s my process for using AI effectively without overwhelming myself:
First, I prompt the AI tool that I want to create a blog based on the outline provided (which I attached as a PDF or copy/pasted in).
Full disclosure: I made myself a GPT that does not require loads of front-end prompting. I just tell it i want to make a blog, load up some attachments, and away I go. What I’m sharing with you is how to blog without a preset GPT.
If you’re using a plain chatbot interface (like what you see when you open Claude, ChatGPT, etc.), include some instructions in your prompt:
I want to create a blog (don’t bother giving it a word count. The way AI chunks data makes it suck at counting.)
I want it based on the provided materials (that’s your outline).
I want to work through it section by section – do not move on to the next section until I have approved it.
As we go through each section, I’ll tell you what I’d like included in the section.
I want X keyword(s) included in each section.
I want you to compile all the sections we create at the end into a single blog.
I want to organize the outline and approve it before we begin drafting each section.
Once that prompt is in, I often ask the AI to help me fill any gaps in the outline from a reader-interest perspective.
If I have too many points, I’ll ask it to suggest which ones could be consolidated or omitted.
If I don’t have enough, I’ll ask for suggestions on what a reader might also want to know.
If it feels like it needs a better flow, I ask it to rearrange the points of the outline into one that make more sense.
Once the AI suggests an outline you like, tell it to move to the next section.
Now, it’ll ask you what you want in the section – or give you a prewritten draft suggestion. Read it. Make sure it has that keyword. Request any changes you want.
Get LONG with those re-prompts. Tell it stories, testimonials, examples, data results, step-by-steps, etc. that you want. Upload documents you feel have relevant information. Original data like this will really juice up your blog. It’s not in the AI database, so it would never come up with it.
If you want to lean more heavily into the AI’s help, request it do research on the topic or sub-topic. Specifically asking it to look at Reddit, Quora, or other forums can give you amazing customer research into what they’re interested in.
Now, if you put in a long re-prompt and the stuff it gives you don’t sound as much “like you” as you want, tell it the magic words: “Adhere more closely to the input I provided.”
Important Note on Adding Your Original Content:
Original content is what truly sets a blog apart for readers AND for algorithms.
An AI-generated blog pulls from its database or information it retrieves online. But it can’t read your mind. Or your private company docs/data. Hopefully.
Your unique stories, client testimonials, case studies, and experiment results add authenticity and relevance.
For example, there are plenty blogs on how to blog with AI, but mine is based on my real experiences and data—not just information AI pulled together from the internet.
OK. Important note complete. Back to drafting.
Once you feel the section is 60 to 90% where you’d like it to be, move through each section until you’re finished.
Why not move on when it feels 100% good?
Because it needs your edits.
And a draft that’s 60% close to done is easier to edit than a draft that doesn’t exist.
At the end, I like to come up with a title, subtitle, meta tag and meta description, all with a keyword or variation of it. Sometimes the chatbot you’re using will come up with a basic title, but I only use it for inspiration.
You know what will capture your niche audience’s attention better than a chatbot would. AI’s knowledge of your niche audience is wide, but your knowledge of it is deep.
You can explicitly ask the chatbot to “suggest a title, subtitle, and meta info including [keyword].” Rework the suggestion, or request more suggestions until you build what you’re looking for.
Ok. Once we’ve gotten a draft, it should feel about 60 to 90% done. It won’t feel like it’s ready to publish. It needs your touch – your unique style of formatting, details you forgot, and sentences reworded to sound like you (or the client).
It’s time to edit.
3. Finalizing Your Blog Outside of AI
Your draft has gotten a lot of invisible labor and editing essentials covered by AI. Proofreading, cleaning up your input and arranging it into cohesive thoughts, basic formatting.
Now the final step in how to write a blog with AI: polish it outside the AI platform.
There’s several reasons for this.
You need to run it through a plagiarism checker. AI is notorious for plagiarizing, which is a huge SEO sucker punch. And AI chatbots are not good plagiarism checkers. Grammarly has a free one. The paid one I like is Quetext.
You forgot stuff. It’s way easier to add it in a doc then to re-prompt the AI a million time to add additions.
You need to add your flair (or your client’s flair). Change things up to match how you’d say it. Delete stuff you’d never say. Add things AI would never know (like original data and stories).
So, once the AI draft is complete, I move it to a Google Doc. This gives me full control over style, grammar, and flow, helping me make it feel genuinely my own.
I review the blog to make sure each section reads smoothly. The easiest (and fastest) way to do it is by reading it aloud. I check for parts that might still sound like AI-generated text and adjust the language to match my style.
Then, I use a plagiarism checker to ensure all content is original.
Once you’ve made all the tweaks you want, run it through a quick proofreader, like Google doc’s built in one (⌘ + alt + X), a Grammarly plugin…or AI! (I’ve got a free one here that won’t tamper with your writing)
Boom. You’re done. That was a bunch of instructions, so I’ll break my AI blogging process down into a concise list in the next part.
POV: Using AI to Write a Blog
Here’s my process when I’m inspired to start blogging with AI:
Capture initial ideas in a Google Doc. Whenever inspiration hits, I open a Google Doc and jot down points I want to cover. Sometimes I do this all at once; other times, I revisit the document to add new ideas.
Do keyword research. I’ll run a related top-ranking blog through AI to find the keywords, use Ubersuggest, or Google a phrase and see how it’s worded in the “people also ask” section.
Upload the outline and fine-tune it with AI. After pulling up my chatbot of choice (I like ChatGPT), I download the outline as a PDF (or copy/paste it) and attach it to the AI prompt:
I ask AI to help me finalize the outline, including suggestions for organizing it more coherently, bonus points to add, items to consolidate, or points that could be omitted.
I make sure my prompt instructs the AI that we’ll work through it section by section, waiting for my approval before moving on. I also include the keywords I want in each section and ask the AI to compile all the sections into a single blog once we’re done.
Draft section by section. With the outline finalized, I work through one section at a time, guiding it to include specific original data I provide, like stories, testimonials, and results. I give the AI this information by typing it within my prompt, using my phone/laptop voice typing, and/or uploading relevant documents. If the blog section feels close to 60 to 90% what I’m looking for, I move to the next section.
Create additional blog details. Once all sections are complete, the AI should compile it as a full blog. I then use the AI’s help to generate ideas for the title, subtitle, meta title, and meta description.
Edit in a Google Doc. I then paste the blog into a document to do my editing and final touches. I also run the blog through a plagiarism checker.
There’s several popular AI blogging “tricks” that I’ve tried and that I choose NOT to use:
Generating images. They look AI-generated, which can imply the blog was AI-written. Nobody wants to read something that nobody bothered to write. Unless it’s an infographic or a super on-brand cover image, I don’t feel AI images add much value to a blog.
Creating fictional testimonials and customer stories. AI likes to slide in fake stories. It’s very convenient, but it’s not honest. Just source real testimonials and stories.
Using AI-generated blog titles. Titling blogs that get clicks takes a really rich (and current) niche understanding. I find AI is helpful for generating ideas for titles, but not coming up with them.
Editing. Editing with AI can be super tricky. Even with detailed prompts, a chatbot often takes out your unique voice with its edits. That’s why I built an Editing Chatbot that simply gives you a suggestion list (and doesn’t tear up your writing).
Let’s wrap it up.
Yes. This blog was written with AI help.
Figuring how to write a blog with AI means finding a balance between fast and original. Through trial and error, I developed a process that lets me create high-quality, SEO-friendly blogs quickly, without losing my unique voice.
In short:
Start with a Simple Outline. Jot down a list of main points in a Google Doc before prompting AI. This outline sets up a structure for a well-organized, readable blog. Plus, it ensures the AI knows exactly what you want.
Draft Section by Section. To avoid a crazy edit process, have AI generate each section separately. Download the outline as a PDF and attach it to your prompt, asking AI to add or consolidate points as needed. This lets you edit one chunk at a time.
Include Your Own Content and SEO. Incorporate personal stories, testimonials, or case studies, along with relevant keywords in each section to make the blog authentic and SEO’d.
Polish the Draft Outside of AI. Transfer the draft to a platform like Google Docs, where you can fine-tune language, add what you forgot, and check for plagiarism. Add final touches like a strong title, headings, meta description, and original images to complete your blog.
AI can enhance your work—but only when you stay in control of the content, voice, and final polish.
A lot of blogging is invisible work. We think about the classic typing-writing part – but writing’s made up of lots of editing, structuring, proofreading, and brainstorming.
Spend your blogging brainpower on the stuff that needs your original touch. Original data and unique edits.
I’m excited to see where AI takes blogging. I really believe the surge of low-quality content will make the higher-value stuff rise to the top.
And of course, I think it’ll make great bloggers put out even better stuff faster.
What do you think about AI blogging?