Over the past few months, I have been excited to start a new phase of learning at this late stage in my life.
My first career choice was as a draftsman. This led me to other opportunities where I eventually became a broadband engineer, designing city-wide cable television networks (when that was a thing in the US). This skill took me to work overseas in Asia, which was the best life choice I ever made.
But when that work ran out, I learned about web technologies, web development, SEO, and content writing. While I still do these tasks, I am intrigued by the latest developments in AI and automation.
I am now learning about using AI and Automation to solve business problems.
When tools like ChatGPT came out, I was energized to learn more, but disappointed by what was capable. That has changed quickly.
As I write this, ChatGPT 4o has recently launched. I have experimented with some of the fascinating features it offers. The automation tools and services available have also grown rapidly in recent months. The two most popular, Make.com and Zapier, have improved their tools significantly to allow AI to integrate with nearly any other service on the Internet. New automation tools are cropping up for a piece of this pie.
In this article, I am going to summarize my initial efforts and some of the things I have tried. But as I dig deeper, my goal is to offer my personal experience as I work through the newest AI and automation tools.
Here are some of the things I have tried.
AI Code Generation Based on an Image
I was working on a website and thought I would give the new ChatGPT 4o a try with image recognition and code generation.
I took the above image and asked ChatGPT to generate the code to make this into a photo gallery. Each image should be clickable and open the image in a lightbox style.
Here is my prompt:
I am going to provide you with an image and I need you to provide me with
the HTML, CSS, and JavaScript code. Here is the image:
[image]Not much I know. Even with such a simplistic prompt, the output was good, but of course, it didn’t exactly match my desired output.
The smaller images were all in one column. They should be in two columns. I pointed this out to the AI and it made the necessary corrections.
The AI was using an updated CSS method display:grid; instead of the older table or div approach. I was happy.
Without going into too much additional detail, it took me about 4 additional prompts to get it perfect.
One prompt was needed to position the “All Photos” button properly over the images. The rest of the prompts were to get the JS code correct. In total, 3 prompts for the HTML/CSS, and 3 prompts for the JavaScript.
I probably could have done everything with one prompt.
If you would like the details just drop me a note and I will post them or send them to you.
I was impressed by what ChatGPT had produced. While this was a standard code snippet I already had, the AI still took about 5 to 10 minutes of work to create an updated version of the code that could be used directly on the website.
Here is the final output.
AI and Automation for Content Generation
This is something that has always intrigued me.
I have tried out a couple of tools that would generate blog posts in bulk, along with images and links, and automatically post them to my WordPress blog.
The first tool I tried was developed by Chris Derenberger at AI Wisemind. The first version launched just after the release of ChatGPT 4 and just before Google’s HCU.
This tool is fantastic at cranking out the content. It uses AI models to generate featured images. You can also insert your Amazon Affiliate ID and the generator will add “Buy Now” buttons at points throughout the content with your affiliate links.
I quickly learned that I needed to review every post to reword it so it sounded more human.
My advice to anyone is don’t use tools to create content in bulk for your blog. Tell your story as a real person and develop your content at your own pace. It will take longer. But it will be valued much more by Google and, more importantly, by your website visitors.
But, I still needed some help from AI to create some of the content. One of the most important areas was with my blog post titles.
Generating Headlines with ChatGPT and Make
I undertook my first automation on Make.com to take the form input from a user, generate headlines, save the output to a Google Doc, and email it to the user.
At the time I was reading an ebook download from Jon Morrow at SmartBlogger on Headline Hacks. There was some good information in there that included some ChatGPT prompts to help make the process easier.
I altered the prompts to better fit my automation. The variables or user input data were extracted and made into form fields in a Google form.
The user would start with the form providing their email address, desired keyword phrase, and their target audience.
That data was then passed to ChatGPT 4 where it was inserted at the correct points in the prompt and processed. The results were saved to a Google document in my Google Drive and then emailed to the address provided by the user.
As a test, I provided the keyword of “Local SEO” with an audience of “Home Service Business Owners”. The output was the desired satisfying outcome for the audience, an image or photo of this outcome, three dominant emotions the audience would feel as a result of the outcome, and five sample headlines.
Here is an example:
Each form submission would produce 5 outcomes each with the same format.
The automation didn’t take me very long to put together. Researching information to set up the prompts was the most time-intensive. But it showed me what is possible with AI and Automation.
Wrapping Up
Currently, I am working on getting my certification from Make.com and hope to be making automation for others to make life easier.
I hope that this new knowledge will help keep me going at this late stage in my career. I have seen older versions of AI that were less than impressive. So, I am enthusiastic about this new wave of artificial intelligence.
