From Blog Posts to Learning Guides: How AI Helped Me Rediscover 15 Years of Content

A couple of weeks ago, I started what I thought would be a simple project. I wanted to clean up my website. There were links to fix, SEO details that needed my attention, and metadata I had overlooked for years. I also wanted to improve the readability of older posts and make sure they had appropriate descriptions for Pinterest.

Nothing particularly exciting. Just the website maintenance that accumulates when you have been blogging for a long time.

But as I started working through my posts, I thought about something entirely different. I was reading articles I had written years ago, revisiting book reviews, and rediscovering ideas I had explored at different points in my career. While some of the material appeared outdated, a large portion remained relevant. In fact, how much useful information I had accumulated pleasantly surprised me.

I remember thinking, “Man, I’ve put together a lot of great content over the years. How do I get more visibility on this stuff?”

That question changed the direction of my project.

What started as a website cleanup has turned into an experiment in using artificial intelligence to organize years of content into learning guides and knowledge briefs. Along the way, I’ve been working with Google Gemini, ChatGPT, and MediaWiki. I’ve also been building AI skills to help me repeat the process without having to start from scratch every time.

I’m still figuring things out, but I wanted to share what I’ve been doing because I think there’s an opportunity here for anyone who has accumulated a body of work.

Rediscovering What I Had Already Created

I started my blog back in 2011 as part of a journaling exercise for a graduate class. Building a knowledge base was definitely not on my mind. I was simply documenting what I was learning and sharing things that interested me.

Over time, that habit grew into something much larger.

Today, my website contains approximately 1,300 blog posts and pages. I’ve written book reviews, shared podcast episodes, documented instructional technology tools, and reflected on ideas related to learning, productivity, and professional development.

Much of this material came from my learning journey. I would read something, experiment with a tool, attend a conference, or explore a new idea, and then write about what I had discovered.

For years, I dedicated most of my attention to crafting the next post. I didn’t spend nearly as much time thinking about how all those individual pieces might fit together.

As I continued my website cleanup, I also noticed something that bothered me. Several older posts referenced learning guides I had created, but those guides no longer existed.

When I worked in higher education, I regularly developed learning guides to help faculty and staff get started with unfamiliar topics. Sometimes, people needed a quick introduction to a tool. Other times, they wanted to explore a subject in greater depth. The guides provided a starting point and pointed readers toward additional resources.

They typically included a combination of materials I had created and resources developed by others.

Seeing those broken references reminded me of how useful those guides had been. It also got me thinking about whether I could bring them back.

Only this time, I had something I didn’t have before: fifteen years of my content to draw from, along with generative AI to help me organize it.

Letting Gemini Look at My Website

Before I could do much with the content, I needed to understand what I actually had.

I used Google Gemini to analyze my website. I exported information about my blog posts and pages into CSV files that included the titles and URLs. Afterwards, I gave those files to Gemini and requested that it examine the content.

What came back impressed me.

Gemini provided a thorough analysis of my website and helped me see how my writing had developed. I could see shifts in my interests, recurring themes, and the different topics I had explored throughout my career.

It was fascinating to see my blog from that perspective. When you’re writing individual posts over many years, you don’t notice the larger patterns that emerge. You’re usually focused on whatever you’re learning or working on at the moment.

I’ll share more about that analysis in another blog post because there were some interesting discoveries.

For now, though, the important part was that Gemini helped me realize I already possessed a substantial collection of material that I could organize into something more useful.

I started wondering what would happen if I asked Gemini to take a particular topic and pull together the related content from my website.

That was the beginning of my learning guide experiment.

What If AI Could Build a Learning Guide from My Existing Content?

Having spent much of my career in instructional design, I consider how people learn and how to organize information to support that learning.

A collection of articles can be useful, but it doesn’t provide someone with a clear path through a subject. A person unfamiliar with a topic may not know where to start or which article to read next.

That’s where learning guides come in.

I wanted to create resources that would introduce a topic, provide some context, and then lead readers through related materials in a logical sequence. Ideally, someone could begin with a basic understanding of a subject and gradually explore more advanced ideas.

And rather than writing everything from scratch, I wanted to use the content I had already created.

Since Gemini had the CSV files containing my website information, I asked it to create an example learning guide using my existing posts.

A few minutes later, I had a guide that included a considerable amount of organized information and references back to my website.

I was thrilled.

I had been considering how much work it would take to search manually through hundreds of posts, identify related articles, organize them, and write the material needed to connect everything. Putting together a single guide could take me hours without any trouble.

Here was Gemini producing a useful starting point in minutes.

Now, I certainly would not take everything it produced and publish it without reviewing it. I still needed to check the references, evaluate the sequence, and make sure the information was accurate.

But the hardest part of many writing projects is getting started. Gemini had taken care of much of that initial work.

I could now spend my time refining the material instead of staring at a blank page.

The next question was where I should put these guides.

Returning to an Old Friend: MediaWiki

As I thought about how to publish the learning guides, I kept coming back to the idea of using a wiki.

I’ve experimented with MediaWiki in the past, and I have always appreciated how easily it allows you to connect related information. If you’re unfamiliar with MediaWiki, it’s the software that powers Wikipedia.

What appealed to me wasn’t simply the ability to create pages. I can already do that on my blog. What interested me was how naturally a wiki supports connections between topics.

For example, suppose I create a learning guide about using dictation as part of a writing workflow. In that guide, I might discuss Zettelkasten to organize ideas. I might also reference Otter.ai as a tool for transcribing spoken notes.

With MediaWiki, I can easily connect those topics through internal links. Someone reading about dictation can follow a link to learn about Zettelkasten, then move to a page about Otter.ai, and perhaps discover another resource about organizing research.

The information becomes interconnected. That was exactly what I wanted. I wasn’t interested in replacing my blog. I wanted to create another way for people to discover and use the information I had already published.

So I installed MediaWiki on my web server.

Naturally, the installation didn’t go perfectly the first time.

I discovered that my hosting environment didn’t have the PHP version required by the MediaWiki release I had selected. After working through that problem, I installed an earlier version that was compatible with my server.

I also returned to FileZilla, a tool I used quite a bit when I was doing more website development work. It had been a while since I needed to transfer files directly to a web server, so that brought back some memories.

I created the necessary database, transferred the MediaWiki files, completed the configuration, and set up my administrative account.

After refreshing my memory about how MediaWiki worked, I had a functioning wiki.

Now, I needed to get some content into it.

Watching Gemini Create MediaWiki Pages

With the wiki running, I returned to my conversation with Gemini.

I had already seen that it could create a learning guide using my existing content. This time, I wanted it to produce the guide in a format that MediaWiki could understand.

MediaWiki uses its own markup language, called wikitext. Rather than manually formatting each page, I asked Gemini to generate the content using the appropriate markup.

And it did.

The result was a central hub page along with five supporting lesson pages. Navigation links connected the pages, and they included references to material from my blog.

I couldn’t have been happier.

I could take the generated content, copy it into MediaWiki, and begin building the learning guide. Although the basic structure was already in place, I still needed to create the individual pages and review the content.

The hub page served as the starting point. It introduced the topic and provided links to the individual lessons. Each lesson explored a particular aspect of the subject and helped move the reader through the material.

I liked what I was seeing.

I’m not committed to having exactly five lessons in every guide. That’s simply the framework I’m experimenting with. As I develop more guides, I may discover that some topics need fewer lessons while others need more.

That’s one thing I enjoy about this kind of project. You don’t have to figure out everything before you begin. You can create something, see how it works, and make improvements.

And that’s exactly what happened next.

From a Good Prompt to a Reusable ChatGPT Skill

As I started working with the learning guides, I noticed things I wanted to change.

Some of the formatting wasn’t quite right. I wanted the navigation to work in a particular way. My goal was to have uniform page titles and an orderly structure for the hub page and the lessons that support it.

Every time I created a guide, I could tell Gemini what to modify, but it didn’t seem like an efficient approach.

My goal was to document all those instructions to reuse later.

That’s when I started thinking about creating a skill in ChatGPT.

If you’re not familiar with AI skills, think of them as reusable sets of instructions for accomplishing a particular task. Rather than explaining your entire process every time you want something done, you create a skill that contains those instructions.

Then, when you need to perform the task again, you call the skill.

I returned to Gemini and asked it to create a prompt based on the learning guide process we had developed together.

I took that prompt into ChatGPT and used it to create my MediaWiki Learning Guide skill.

What I found useful was how easy it was to refine the skill.

For example, I could tell ChatGPT that I wanted a specific navigation template at the beginning of every lesson. I could instruct it to format the links differently or change the way page headings looked.

Instead of rewriting the instructions myself, I could simply explain what I wanted changed and ask ChatGPT to update the skill.

Then, I would test it. Sometimes, the results were exactly what I wanted. Other times, I discovered something else that needed attention. I would provide additional instructions, update the skill, and test it again.

I spent about an hour going back and forth, making minor adjustments and gradually improving the output.

The skill still isn’t perfect, but it’s getting better.

Keeping Source Content Available

One challenge I’m working through is how to make my CSV files consistently available to the skill. Those files are important because I want the AI to look at my existing content before creating a guide.

I asked ChatGPT how best to handle that, and it suggested using a project to keep the source files and related materials together. That’s another piece of the workflow I’m exploring.

I’ve also discovered that I can download the skills I create. I enjoy having that option because it allows me to preserve the instructions and potentially share them with other people.

What strikes me about this entire experience is how my approach to working with AI is changing.

Not that long ago, I would ask an AI tool to help me write a prompt. Now, I’m asking it to create a skill. It’s a minor change in wording, but a significant change in how I think about getting work done.

Not Everything Needs a Full Learning Guide

As I continued experimenting, another idea occurred to me.

Not every topic needs a complete learning guide.

Sometimes, I just want to provide someone with a quick overview of a tool or concept. I want them to understand what it is, why it might be useful, and how to get started. I also want to point them toward additional resources if they decide to explore further.

That’s where the idea of a one-page knowledge brief came from.

For example, suppose someone wants to learn about Otter.ai. They may not need five lessons explaining every aspect of the tool. They might simply want to know what Otter.ai does, how it works, and some practical ways to use it.

A knowledge brief can provide that information in a concise format.

I also want these briefs to reference material I have already created. If I have written about a tool, reviewed a related book, or discussed the concept in a podcast episode, I want to include those resources.

So I developed another framework and created a second ChatGPT skill, this one focused on generating MediaWiki knowledge briefs.

I’m still testing and refining it, but I like where it’s going.

The learning guides and knowledge briefs serve different purposes, yet they complement one another.

A knowledge brief can introduce someone to a topic. If they want to go deeper, they can follow a link to a learning guide. From there, they can explore related topics and discover additional resources on my blog.

The more I think about it, the more possibilities I see.

Bringing Everything Back to My Website Cleanup

What’s interesting about this project is that it hasn’t replaced my original website cleanup effort.

Instead, it’s become part of it.

I’m still working through my older blog posts, improving the information and making them easier to discover.

To help with that process, I’ve developed four AI skills that I use regularly.

One helps me identify an appropriate SEO focus key phrase. Another reviews a post for readability and recommends where I should add headings, break up paragraphs, or improve awkward sentences.

A third creates meta descriptions, something I didn’t pay nearly enough attention to when I originally published many of my articles.

The fourth creates Pinterest descriptions, so that the posts are better prepared for sharing.

These skills have made the cleanup process much more manageable.

Now, as I’m reviewing a post, I can also think about how it might fit into the larger knowledge base.

If I come across a subject that deserves a learning guide, I can use my learning guide skill to create one.

If I encounter a tool or concept that would benefit from a quick reference page, I can use the knowledge brief skill.

I’m essentially working on two projects at the same time.

One is improving the blog itself. The other is creating a new way to organize and share the information contained in that blog.

And the nice thing is that the two projects support one another.

As I improve the older posts, they become better resources for the learning guides. As I create learning guides, I give readers new ways to discover those older posts.

I still have a long way to go. With approximately 1,300 posts and pages, this isn’t something I’m going to finish in a weekend.

But I now have a process that makes the work much easier.

What This Experience Has Taught Me About AI

I’ve been experimenting with generative AI for quite some time, but this project has reinforced something I think is especially important.

AI becomes much more useful when you stop treating every interaction as an isolated request.

I could have asked Gemini to create a learning guide, copied the result, and moved on. That would have been useful, but I would have missed a larger opportunity.

Instead, I kept asking questions.

How could I improve the structure? What steps can I take to make the formatting more consistent? How could I reuse the process? How could I connect the guides to my existing content? What other resources could I create using a similar approach?

Each question led to another experiment. And each experiment helped me improve the system.

I also recognize that AI isn’t doing all the work for me.

My background in instructional design helps me determine whether the learning sequence makes sense. My experience with websites helps me understand how the content should be organized. And my knowledge of the material helps me evaluate whether the AI has produced something accurate and useful.

I still need to review the output. I need to verify links, check references, and make sure the information reflects what I want to communicate.

But I don’t have to do all the initial searching, organizing, and drafting myself.

That’s a tremendous advantage.

Applying This to Instructional Design

For those of you working in instructional design or talent development, I think there’s a genuine opportunity here. Many of us have developed processes we repeat regularly. We create lesson plans, learning guides, job aids, reference sheets, and other instructional materials.

If we can capture those processes in reusable AI skills, we can reduce the time spent on repetitive work and focus more attention on the quality of the learning experience.

And you don’t need to use MediaWiki to take advantage of this approach.

MediaWiki is the tool I’ve selected because it fits what I’m trying to accomplish. You might use a different platform entirely.

The important part is thinking about the workflow and how AI can help you make it more efficient.

Where Do I Go from Here?

Right now, I’m continuing to work through my blog posts.

As I clean them up, I’m identifying topics that would make good learning guides or knowledge briefs. I’m also continuing to refine the skills so they produce results that are closer to what I want.

Eventually, I’d like to have a substantial collection of interconnected learning resources.

Imagine someone visiting my website because they’re interested in Zettelkasten. They find a knowledge brief that introduces the concept. From there, they discover a learning guide that walks them through the process in more detail.

As they explore, they encounter links to related topics such as note-taking, research, writing, and transcription.

Perhaps they discover an older blog post they never would have found otherwise.

That’s what I’m trying to accomplish.

I want the information I’ve created over the years to be more useful to other people.

And frankly, I want it to be more useful to me as well.

I’ve spent a considerable amount of time reading, experimenting, writing, and sharing what I’ve learned. It seems a shame to have all that material scattered across individual posts without a better way to connect it.

The wiki gives me a place to bring those ideas together. AI gives me a way to do it without spending countless hours manually organizing everything.

I’m sure the process will continue to develop. I’ll probably change the learning guide framework, adjust the knowledge briefs, and discover additional ways to use AI skills.

That’s fine with me. I enjoy figuring these things out.

Final Thoughts

When I started cleaning up my website a couple of weeks ago, I had no intention of building a wiki or creating a collection of AI skills.

I just wanted to fix some broken links and improve my SEO.

But that’s one thing I enjoy about learning and experimenting with technology. Sometimes, you work on one problem and discover an opportunity you hadn’t considered.

If you’ve been blogging, teaching, developing training, or creating resources for any length of time, I encourage you to take another look at what you’ve accumulated.

You may have more useful material than you realize.

Maybe you have presentations that could become learning guides. Perhaps you have written articles you could organize into a reference library. You might have training materials you could transform into job aids or knowledge briefs.

The tools are available to help you make those connections.

You just need to experiment.

For me, this project has changed how I look at my website. Instead of seeing 1,300 individual posts and pages, I’m seeing an interconnected body of knowledge that I can continue building upon.

And I think that’s pretty exciting.

Here is the beginning of my wiki.

What about you? Have you used AI to bring new life to content you’ve already created? I’d love to hear what you’re doing.

Leave a Reply