When I published my first blog post about a year ago, I was pretty green. Looking back on this moment now, it was as much a lesson as it was a milestone.
It was in my first week of immersion in AI that I heard somebody say:
Start writing and don’t stop.
The benchmark I remember was about 2 hrs. a day which sounds about right now that I’ve got a few solid pieces of content under my belt.
A few months ago, I stopped writing and reset my content goals. If I want to produce content that enhances rather than questions my authenticity, it needs purpose and context.
Only on a solid base of publicly scrutinized content will the internet favorably determine your level of trust. As a first impression, the purpose of your content as a part of your brand should be clear, sharp, got it – let’s read some more.
My content plan is still less of a plan than some form of commentary aligned with my build roadmap – content is one way to draw attention to commercially driven activity, especially if it has authenticity, I presume.
Bootstrapping the AI evolution
Back then when I published “bootstrapping the AI evolution”, my first public writing, it was on the back of my first big milestone as a freelancer, building a professional public website.
Although I am an avid follower of several freelance programs such as Liam Ottley AI Accelerator, all of which insisted in not getting caught up with too much infrastructure, I just don’t have it in me to put myself out there with anything other than my best foot forward.
The success of Liams and other programs can’t be ignored but their own success is predicated on content and lots of it. I knew this back then, but what I didn’t know was how they did it.
What I observed was that the most engaging commentators didn’t produce a lot of content, rather they re-purposed content themes in many different ways. The difference is that it’s not linear. I noticed content being reposted, with intent, and content snippets filling gaps that lead to new content.
Most of all I noticed that the content had authority. I believed it was made from an authentic source. There’s got to be a subliminal element with authenticity, it’s sometimes just a feeling.
I was genuinely impressed, especially with Liam Ottley who has built as a young entrepreneur, a formidable online community, a YouTube following of 800k and an AI centric software company going gangbusters. Ok, Liam is a fellow kiwi but this guy’s more popular than most Euro League Soccer champions.
Liam Ottley became a genuine hero for me once I saw his machine in action. At the root of his success, was authentic content supporting, complimenting and increasing the feeling of trust that people get when content makes them sit and say: “That’s a great idea. I can do that! I will do that!”
I knew however, that it had to involve automation. The amount and consistency along with post maintenance on likes and hundreds of comments, it was just too much for one person to do any justice.
This left me thinking that maybe you realistically had to decide – am I a content creator or am I a technology provider – you can’t be both.
Creator or Practitioner: A choice I refused to make
Writing content is and should never be easy. Even the greatest writers experienced writers’ block, a slowdown of creativity or some time when they couldn’t find the right words or even the right subject. That’s because at its heart, writing is a creative endeavor, it comes from the soul, and the soul only produces the best writing once you’ve earned it.
For me, writing content has always been a process consisting of layer upon layer of details that fall together over a period of time. Usually started with a thought, some are committed to an ideas scratchpad to see the light another day, while others skip that step instead being wrangled into a cohesive argument over several days.
But good writing can only come with thought.
This is the ramblings of your mind we’re talking about here, it’s not insignificant and if you’re anything like I am there’s no guarantee it will make sense!
To produce authentic content, you have to go through some pain to earn it. It doesn’t pour out the same way for a guy in an office prompting AI than it does an authentic writer with head in his hands in frustration.
Producing just one blog post demonstrated to me that it wasn’t sustainable as a side gig without automation. Interestingly enough, publishing a simple blog post turned out to be one of the more challenging aspects of having an online presence I came up against.
Either way, I wasn’t about to put blogging down to experience just yet.
The Perils of Content Creation at Scale
Writing and posting an article in different formats to suit different channels, manually, requires a lot of manually steps and a tracking mechanism such as a spreadsheet.
Most people tend to write for one channel and re-purpose from this base. I’m no different. My preferred channel is LinkedIn because it’s one channel I’ve consistently maintained over my career.
Facebook has traditionally been personal for me so now I have a page to separate out my commercial concerns from my private. This is where you can see an immediate need for at least reviewing the content for fit before cross posting from LinkedIn.
I’m already 2-3 days in at this point, not solid editing time rather timespan. I’ve written an article and posted it to LinkedIn. But it’s not:
Posted to my website blog section
Sent it to my subscriber base
Re-formatted for other channels
Posted to other channels
What does content automation look like
To automate content with the help of AI is a slippery slope. If I had it my way, there would be similar barriers to that which LinkedIn have placed on their API that prevents bots from posting on someone’s behalf.
But the big mistake to make automating content is not where and how often you post, so much as it Is, what you allow AI to do with your content before posting, assuming you start with your own content in the first place.
Establishing Moral Boundaries
After a long road of stops and starts, I’ve finally managed to build a content pipeline that semi-automates most of the steps mentioned above. But AI rarely makes an appearance above my database layer where I use it sparingly to:
Register key data such as title, key words and links to other content
Reformat base article into short form text for other channels using predefined, deterministic rule sets and skills
Capture gate markers and notify me for review
Track engagement once posted
However, my design precludes AI from rewriting my content and each new iteration is scored on how far it deviates from the original.
Herein lies my personal #1 rule about content:
Don’t use AI to create content that you author and put into the public arena.
My pipeline is not fully automated. I have to hit a publish button to post anything into a channel and some channels, LinkedIn case and point, don’t allow auto posting in most circumstances.
There’s no arguing that this creates some efficiencies for me, however, it takes a lot of trial and error building it. In terms of cost benefit, I would say my first working prototype took almost as long to complete the process as it does manually.
A prototype is just the beginning of a process that I’ll be fine tuning for a long time to come.
The challenge for me is that I rarely use AI generated text for my own writing because it doesn’t sound or feel like I’ve written it.
Automating Content – Is It worth It?
This is the question I’ve asked myself many times as I churn through update after update to my Supabase Edge functions. These database functions can use AI to transform incoming data but, to transform written text using AI, you must appreciate the concept of deterministic events.
The less structure you enforce on AI when generating content, the more unpredictable the outcome.
It makes sense then, that AI is best used somewhere close to the point where it’s possible to prescribe the format, style, tone and subject matter of a single content piece.
It is possible to create predictable content using AI, but some use cases are more suited than others which is why my own experience attempting to automate it, has seen many iterations. My content is creative rather than consistently structured hence not repeatable.
Long before I even set out to attempt such automation, I knew it would always look like this. I’ve seen as many structured reports and documents as I have creative reads, and you’ll never automate creativity without giving up creativity itself.
Structured reporting and other repeatable content on the other hand can use coding techniques to steer the AI engine. Techniques such as task splitting, parsing, json schemas and others are nothing new but provide a route for the AI to follow.
Authenticity – The Foundation for Trust
While there are many techniques to induce AI to produce very good content on your behalf, author written content will always trump AI generated content when it comes to authenticity and trust.
Call it a robot thing because sometimes it just doesn’t feel right or even a romantic notion, but I feel honest and clean when I go to an effort to write content.
I don’t have to question my moral compass, my integrity is intact and it is only on this basis that you can develop true authenticity.
Only from authenticity can you build trust.
Mike G Robinson
Mike develops deterministic, event-driven AI systems focused on workflow automation, operational intelligence, and enterprise process orchestration. With a background in Accounting, IT Risk Management, SAP FI, and enterprise transformation across Australia and New Zealand, his work centres on practical AI implementation that improves decision-making, governance, and operational efficiency. Mikes’ Profile
See my Linkedin Company page: MGRNZ
See my website: MGRNZ.com
Go to My Services Page: What I’m Building
If you’re interested in safely adopting AI in your business, book a 20-minute chat with Mike.
Leave a Reply