AI Detection Part 3

Hey,

Welcome to Part III of our "Avoiding AI Detection" series.

Previously, we covered AI triggers like vocabulary and paragraph structure.

Today, we’re tackling another sneaky factor that gives AI-generated content away: your data sources. (Or really, lack thereof).

AI-written content can lack depth because it's simply adding (what it calculates to be) the most appropriate next word to a sequence of words.

That’s why a key ingredient to avoiding detection (and writing stronger content) is to incorporate real sources from which the LLM can draw content.
 

Here’s the game plan:
 
Step 1: Find solid data sources (Perplexity is your best friend)
 
Don't bother with a Google search... Go straight to Perplexity.ai and get one (or multiple) sourced articles for ChatGPT (or Claude or DeepSeek) to reference as context. 
 
Ask Perplexity for:
Recent reports & studies (Example: "What are the latest statistics on email marketing conversion rates?")
Primary sources (government websites, university research, industry white papers)
Unique expert insights (Substack articles, niche blogs, or original research)
Here's a sample prompt for Perplexity.AI:
I am writing an article on [FILL IN THE ARTICLE CONCEPT].  Please find me three highly relevant sources to fully inform this topic with credible data.  For each of the three relevant sources: provide a summary of the article, provide the names of the author(s), provide the key data points from the source, and provide the article URL.

When it comes to AI Detection, the better the input, the stronger (and more undetectable) your AI-generated content becomes.
 
Step 2: Guide AI with your sources
 
Once you have strong data sources, tell the LLM to use it.  
 
Here’s how:
If you're using the CRAFT method:
(C)ontext
(R)ole
(A)ction
(F)ormat
(T)arget
 
Drop this into your ACTION section:
In the article, please reference the following data sources:
- [Data Source 1]
- [Data Source 2]
- [Data Source 3]
Make sure to reference them naturally and include the relevant statistics provided.
 
When you provide sources, AI can’t default to vague or outdated information.
 
This makes your content more original and harder to flag.
 
Step 3: Make it sound human
 
Finally, tweak AI’s output:
Add personal opinions or experiences (Something AI can’t do)
Use storytelling or examples from real-world scenarios
Pay particular attention to the introduction and conclusion sections of the article.  Adding personal touches in these sections will go a long way in avoiding AI detection.
The result?
 Engaging, well-researched content that flies under the AI detection radar.

Final thoughts
 
If you use this method in addition to vocabulary exclusions and structural changes – you’ll have the golden trifecta of avoiding AI detection.
 
Give it a try and let us know how it goes.
 
Talk soon,
Brian & Andrea

P.S. We're planning an online learning session on 
February 27th (10:00am Pacific / 1:00pm Eastern).

This will be focused on creating Custom GPTs for your business:
-Write articles, emails, and social media in a consistent brand voice;
-Create sales proposals customized for your prospect without recreating the wheel;
-Develop a targeted social media calendar each month with minimal effort.

Potential course titles include:
"Create a Custom GPT that will make your mother proud" 
~OR~
"Learn to build Custom GPTs so good your high school sweetheart will forever regret dumping you"
 
There will be a fee to attend ($20).  

BUT, the first twenty people to reply to this email with "Interested" will get a coupon code that will cut the cost to $10.

HURRY!