AI Blog-Writing: Learnings From My Hour Autonomously Writing Content For MSPs

I want to start this article by saying that I have never been a fan of “spinning” content. In my opinion, this represents all that is lazy and wrong about creating content for the internet and it often leads to poor quality fluff that does nothing but get in the way of a reader looking for a helpful answer. If you are unsure of what I am talking about, article spinning is basically an autonomous way of creating new content from something that already exists. It rearranges sentences, replaces words, and alters the content just enough to make it appear original.

When I started to look into the use of AI to write IT-related content, I assumed that this is all I was going to find. I figured that the same companies that developed these article spinning tools years ago for SEO, were now just using the “artificial intelligence” term to make this seem like something new and interesting. I will gladly admit that I was wrong. What I found is something truly remarkable and I haven’t been able to stop thinking about it since. Here is a little more information on this experiment and what I learned: 

Creating Content From Turnkey Ideas 

Inside MSP Spark, we have over 200 IT-related blog article ideas just waiting to be written (with more added every week). While this is a great resource for our members, it’s only useful if they can execute on these ideas. Some do this by outsourcing them to an agency, others pass them to an in-house employee, but those who get the most value from the resource have some sort of process in place to do so. 

Since these additional processes are critical for our member’s success, I am always looking for new and innovative ways to solve these problems. Recently, after reading about some of the advancements in AI written content, I decided to finally give this a shot. Worst-case scenario was that the writing would be as bad as I had always expected it to be (and I could get back on my high-horse to tell everyone why thoughtful and original writing is so important).

Examples Of Content Written W/ AI 

So begins my test. My first task was to select an idea from our database to use in this experiment. Since I wanted to give this a fair shot, I took the article’s structure into consideration. I thought either “lists” or “comparisons” would probably have the highest chances at success, primarily because the likelihood of this content already existing in other forms was pretty high. So after a few scrolls through our platform I decided on “The Difference Between A PCI Compliance Audit & A Scan.” 

With a simple copy and paste of this title into my AI tool, I was able to generate an intro paragraph instantly summarizing this concept: 

Compliance audits and compliance scans are both important parts of a company’s PCI compliance strategy, but they serve different purposes. A compliance audit is the most in-depth type of security assessment conducted by an external party to determine if your business is compliant with the Payment Card Industry Data Security Standards (PCI DSS). A scan, on the other hand, typically takes less than two hours and can be done internally or externally. The purpose of a scan is to identify any vulnerabilities that exist within your network before attackers take advantage.

As I read this, I couldn’t believe the nuances in the language. It interpreted my title exactly as I hoped it would and provided a brief summary with more precision and detail than I ever could as a writer. Is it a little dry? Yes, but this is also because I chose not to adjust the “tone” of the writing in any way because I didn’t want to skew my initial results.

From this point I decided to outline the rest of the article manually by writing in subheadings and letting the system fill in the blanks. I wanted to make sure that the system was aware of who the audience was supposed to be as well as who was writing it. I did this by offering clues within the H2 subheadings such as: “Finding An IT Provider That Knows PCI Compliance.” The system picked up on this immediately, and as an example, what it returned for this section was almost exactly what I intended: 

If you are unsure whether or not your IT provider is PCI compliant, ask them about their security processes. Any reputable company should be able to provide information regarding how they protect data and what services they offer in comparison with the costs of doing so. If you have any further questions, contact a representative from an external compliance audit firm that specializes in assessing small businesses’ needs to learn more about which additional steps may need to be taken based on where your business stands when it comes to online security.

Your primary goal through completing PCI compliance audits and scans is protecting customer’s sensitive financial information by minimizing vulnerabilities within your network before attackers take advantage of them for malicious purposes like stealing credit card details or identities. This way you can feel confident knowing that all angles have been covered so you can focus on running your business while letting an IT provider handle all of the security oversight.

Problem-Areas That I Experienced 

While I was completely impressed with my results so far, I will say that the results (and the process) was not perfect. While it does seem like “magic” at times, there are other times when you realize that this thing has a mind of its own. Here are a few areas that caused problems for me the first time around that definitely impacted my results: 

• I found it interesting that the system mentioned actual products and brands several times within these articles. In the case of this experiment, it originally referenced several different PCI Compliance scanning tools. While this is not necessarily a negative thing, it’s not a result I was hoping for and thus I had to do some manual editing to scrub these mentions out. 
• There is a definite learning curve to writing this way and it’s not the same as just spinning an article with a click of a button. You have to command the system to get the results you are looking for and that’s not easy to do. Like I mentioned before, it truly has a mind of its own and if you simply just let it write whatever it wants, it’s going to take you in a direction that you may not have intended to go. 
• I found it difficult to produce articles longer than 800 words, as the system would start to repeat itself (not verbatim, but repeating concepts). This is an area that I think could be improved upon with a better outline, but it is an example of the “wide” rather than “deep” disposition that the system has. 
• There are no “new ideas” that come out of content written with AI. The system is simply researching millions of pieces of content on the web and reporting back what it finds in its own unique way. If you are looking to create content that is more thought-provoking, this is not a good solution. 

Why I Think This Could Help MSPs 

While I still have a lot to explore in the area of AI content creation, I am now dedicating more time to this project to find out how we can help our members better leverage this technology as a possible solution. The potential is very real and I personally will be taking it more seriously from this point forward. There are several reasons why I think AI could be a great solution for some MSPs that are struggling with a written content strategy: 

• It’s cost-effective – Not all MSPs can afford to pay a freelancer $100-$300 per article and most agencies require a recurring monthly contract. 
• It saves time – If you are writing content yourself and get no enjoyment from that activity, then this could be a good way to shorten the time spent on these tasks. Also if you are paying an employee to write for you, AI will reduce their time required to do so. 
• It’s well-researched – One of the challenges of writing about IT services is that it requires some level of “technical research.” Sometimes freelancers don’t take the time to do this and end up producing content that is technically incorrect. This solved that problem in my opinion. 
• It’s sustainable – The reason that most written content strategies fail is because they simply were never sustainable to begin with. Either they rely on someone that is not a permanent member of the team, or they rely on time / money that you may have now but won’t have later. AI allows you to create a more sustainable process which can be infinitely repeated.