Week 2 - Group 1 Discussion

What were three changes that you made?

1. I added AI-generated images throughout the lesson blocks. I used the existing block text as the starting point for each prompt, then generated images in the Vector Blues style to create a consistent visual look across the course.
2. I used an AI text prompt to emphasize key terms by bolding them and changing the bolded font color to #0e497b. This helped make important words and phrases stand out consistently across the blocks.
3. I used AI to create knowledge checks at key points throughout the lessons. For each knowledge check, I also created a banner that included an AI-generated image, a heading, and a brief lead-in statement.

Which AI feature or capability was most valuable in your workflow, and why? Did anything about it surprise you?  

I found the AI image generator especially valuable because it can save a significant amount of time when searching for appropriate visuals. I used the Vector Blues style throughout this project, which gave the images a consistent look and helped me explore a style I would not typically use in my usual workflow due to my employer’s design standards.

I also found the text-editing prompts useful, especially the ability to bold key terms and apply a specific font color. That is something I currently do manually in my courses, so being able to automate it could save time and improve consistency. I used the key takeaway feature several times and was surprised by how helpful and usable the results were. The knowledge check generator was also effective, and I can see myself continuing to use it as a starting point for future lessons.

I added audio to the lessons using the AI Audio feature which I liked a lot. It made the VO generation process a lot quicker and easier. The quality of the voices seem good and probably better than WellSaid Studio we’ve been using in my company.

What challenges did you face when trying to get AI to produce the results you wanted? 

One challenge I ran into was with a flashcard block that I had created using the instant-change feature. I liked the initial result, but when I tried to edit the individual cards, the block became unresponsive. I could not select the text or make changes to that block, or to the blocks immediately above and below it. I tried to exit the course, but at first, it appeared the course might not save and that I could lose my changes. Eventually, the application seemed to catch up, and I was able to exit and reopen the course, return to the affected blocks, and make the edits.

I also noticed that prompts for the AI image generator can become very lengthy. In some cases, it seemed like a shorter, more focused prompt might produce better results. I think it will take some practice to find the right balance between giving AI enough direction and giving it too much information.

One question I still have is how accurately Magic Text import can generate course content when the script, language, and activities have already been defined in advance. In my workflow, I typically use a detailed design document that has been reviewed and approved by subject matter experts and management. If that document includes specific language and approved activities that need to be preserved, I want to understand how well AI can use that predefined content to create a new course without changing the intended wording, structure, or instructional approach.

Link: 

https://share.articulate.com/VRUvVX6mnnEO95OqQd9or