How I use AI in my blog writing services without losing the voice (or feeling guilty)

If you’ve been on LinkedIn lately, you’ve probably seen a lot of posts about AI detectors.

Clients are rejecting content because:

• An AI detection tool said it’s 100% AI-generated.  

• There are too many em dashes.

• It consists of commonly used AI words and structures.

At one point, after reading all this, I began to feel that using AI is a crime. 

So I looked into how AI works and learned to use it without losing my tone or the client’s voice in the content.

Let me take you through how I use AI in my content writing workflow. 

But first, let's look at whether using AI for blog writing is ethical.

Is it ethical to use AI for blog writing?

From the SEO perspective, it’s completely fine to use AI-generated content. 

Google does not penalise 100% AI-generated content. 

However, if you want to rank on the search results, you need to follow the Google E-E-A-T framework.

Additionally, Ahrefs found that 87.8% of Google’s AI Overviews use content written by both humans and AI.

Source: Ahrefs

So, there's nothing wrong with using AI to some extent in your content.

What’s my view on using AI blog writing tools?

There are several AI tools for blog writing like ChatGPT, Jasper, Surfer SEO, Frase, and more. 

They help with:

• Topic ideation

• Structuring

• SEO keyword research and planning

• Editing

• Research 

  And more. 

However, human input is vital to make your content stand out. If you use a vague prompt, the AI-generated content will lack tone and originality. For example, the screenshot below is of a website that has used 100% AI-generated content. 

When you read the content in the screenshot, you quickly understand it’s AI, and you lose trust. 

The content mostly focuses on the company’s greatness. It fails to highlight the pain point or connect with the target audience. 

So, if you’re using AI to create content, make sure to use smart prompts and edit the output to sound more human.

How I use AI in Blog Writing 

Since 2022, I’ve tried out several AI tools to help with writing and editing content. And here’s how I use it in different phases.

Phase 1: Blog topic generation using AI

When I need to finalise a blog topic, this is what I do.

First, I figure out the search intent, get details from the client, and pick a primary keyword using a keyword tool.

Let’s work with the keyword “SEO tools for the MarTech industry.” I search the keyword and note the blog titles from AI overview, ads, and top results. Additionally, I also note some of the questions from "People also ask" and keywords from "People also search for." Next, I input all this data into an AI tool to generate a few blog ideas. 

Here are the key things you need to add in the prompt:

1. Primary keyword

2. Semantic keywords

3. Buyer persona details

4. Location

5. Ranked, sponsored, and AI overview blog titles

6. “People also ask” questions 

7. Marketing funnel stage: TOFU, MOFU & BOFU

8. Type of blog (How-to, Listicle…)

Pro Tip: Ask AI for justification for each of the blog titles it generates. 

Phase 2: Using AI for blog outline & identifying gaps

I study all the blogs that are already ranking on the first page of search results.  

This includes: 

1. Blog Title

2. Intro

3. Use of H2, H3 and so on

4. Keywords usage 

5. Tone

6. Number of images used

7. Stats and data used

8. Backlinks and interlinks

This helps me identify all the gaps I can fill in. I do this manually, though. However, you can use AI as well. You have to input the competitor blog content and ask it to identify the gaps. 

You may also use SEO tools such as Semrush or Ahrefs to identify the SEO gaps.  Next, I ask ChatGPT or Gemini to give an outline on the title selected. To demonstrate, I’ll use the blog title, “10 Ways X Can Enhance Your MarTech Strategy”. 

I used Gemini for the blog outline. Here’s my prompt: 

“Hey Gemini, I want to write a 2000-word blog on this topic: 10 Ways X Can Enhance Your MarTech Strategy. It's targeted at UK marketers who are working in the MarTech industry. Could you please give me a blog outline that is SEO friendly, and the tone should be friendly and educational.”

And here’s what Gemini generated. 

Now, I don’t follow the exact structure given by the AI. I use it as a guide. 

To create my blog outline, I refer to the AI blog outline along with the gaps I have identified. I also include some of the stuff that’s working well for the competitor blog. The idea is simple: find gaps in top blogs and give your audience what’s missing.

Phase 3: Draft blog content with AI

I start filling in the blog outline I’ve created. This is mostly manual work. 

I plan and draft the angle, the tone, and the differentiator. While working on this, I’m also mindful of using the SEO keywords, interlinks, external links, FAQs, and more. 

For each subhead, I try to add unique elements like stats, diagrams, quotes, or social media posts if it makes sense.

I use Perplexity AI to find these statistics and data. However, I still prefer doing traditional online research. I explore research papers, expert blogs, and studies to find the data I need to back my point. If the topic needs deeper insights or strong data, I talk to subject matter experts. I also refer to a few points mentioned in top-ranked blogs. If they’re solid, I reframe and improve them in my own words.

So, I use: AI + competitor blogs information + my research to create the content. AI is useful to get rid of writer’s block, especially when you have to fulfill an urgent deadline. 

If I get stuck, it’s usually at the intro. That’s when I turn to ChatGPT to generate a few variations around the same concept to get things flowing. I mostly use AI to get ideas, dig into research, or make sense of jargon I come across.

Phase 4: Clean and edit blog with AI

The best part about using AI for blog writing? It helps in editing. 

I ask it to find grammar, punctuation, and sentence structure mistakes so I can fix them before publishing. I also ask for suggestions on how to improve it. 

But, I don’t follow everything blindly. I consider AI’s edits and suggestions only if they fit perfectly in the content. 

Other apps that I use for editing:

• Grammarly: For those punctuation, grammar, and spelling fixes 

• Hemingway Editor: To fine-tune for readability. 

In my experience, using AI and these tools together seriously levels up your content quality. For example, if Hemingway Editor flags a sentence as “very hard to read”, I use an LLM tool to simplify it. But I don’t just copy-paste. I always adjust it to suit my tone, so it still sounds like me or the client. 

How I maintain brand voice while using AI

One thing to keep in mind while using AI in blog writing: vague prompts = dull, robotic output. However, even when the prompt is specific, it doesn’t generate the same tone and voice you expect it to. 

That’s why I write the content myself to keep the brand voice clear and real. I study the client’s brand guidelines and past content to understand the voice. 

I look into their:

• Sentence style

• Tone

• Commonly used words

• Metaphor uses

• Cultural context

Using these data, I write the initial piece. I also check the client’s past blogs or posts to match their tone.

In my first draft, if I am not satisfied with some of the paragraphs or sentences, I run them through an LLM tool. I prompt it to maintain the same style of writing but improve it. In this way, I don’t need to give any voice guidelines in the prompt. It studies the writing style and refines it. 

But don’t drop your whole blog into an LLM. Just use it for the parts that need refining or when you’re completely out of ideas.

I use this technique mostly for my LinkedIn posts. Sometimes, when I need to write fast or on new topics, I use AI. For instance, one of my past clients was a clinical psychologist. While I handled the content ideation based on trends, I used ChatGPT to draft most of the posts. Here, I generally give a voice and tone-related prompt.

After the content is generated, I edit and tweak it to match the brand’s tone. 

My guidelines for ethical AI use in blog writing

Working as a freelancer or solopreneur doesn’t mean working without guidelines.

Here are a few AI use rules I follow: 

1. Stay clear of 100% AI-generated content

   Only use AI for brainstorming, researching, and gathering ideas. Always tweak content if it’s AI-generated. 
   

2. Disclose use of AI

   I do not hide the fact that I take the help of AI to research and improve my content. I inform my client if I’m using AI in any of my processes. 
   

3. Confirm every detail

   AI models, including ChatGPT, are known to hallucinate facts. So, I cross-check data just like I did before the AI era. I don’t include any information in my content unless I’m confident it’s backed by original research.
   

I’m a Content Marketer, and AI only assists me

AI in blog writing will continue to be discussed on the internet for a while. But AI can never replace human writers. 

Besides, Google prefers AI + human-generated content. So, content writers who use AI smartly will definitely have the edge. 

Find this as informed and incisive as we did? Follow Ambika on LinkedIn and X or comment below to let her know!