SaaS products may require videos as an way to to convey the use of a product to a learner in a way that doesn't overload the cognitive side of the brain. However, developing these videos comes at a cost, and often times the business wants these to be developed rather sooner than later. In this project, I already had many lab instructions and the mandate was very simple : "Take the lab instructions and convert these lab instructions into videos for learners to consume". The problem however, is that converting these instructions into a workable script is a time consuming task. Click on the Process tab to learn how I approached this problem.
Lead Learning Experience Designer
In this challenge, an LLM such as Claude can be used to speed up the process by quite a bit. By creating a skill and defining a proper pipeline, I've built a method that allows any content developer to quickly generate instructional design laced videos based on virtual lab instructions. Take a look at the steps below, describing my approach to resolving this problem.
01 · Export Skillable Labs
First, the Skillable lab instructions and images need to be exported, so the LLM can ground its output in the information that is required.
02 · Invoke Lab-to-Script Skill
Attach the Skillable export to a new Claude session and the skill analyses the lab, then writes the script as a Word document.
Each recording step has two parts: Action, what the developer records on screen, and Voice over, what the narration says. The skill also sets the inclusive language and how the script refers to what's on screen.
03 · Review Script & Record
Once the script is ready, it needs to be reviewed and slightly adjusted. Usually the script gets a developer 90% of the way there. And then everything moves into the next stage of the video production pipeline and everything can be recorded.
The development time of a workable script reduced by 90%, and the overall development time of a video was cut down by half. That's a significant decrease. Once the script is ready, it will then be used for the actual production of the video. Because the entire script is already in text format, it can also be easily used to create closed captions and if the script is combined with an AI voice tool such as Audiate or Elevenlabs, updating sentences becomes very easy and fast.
I selected this project for my portfolio, because the output is fast and efficient, it is also consistent. The production judgment lives in the skill, a newer developer now produces the same demo-not-tutorial, translation-ready script like how an experienced one would, so the saving comes with a raised quality floor rather than a traded-off one. The skill has been in production with the content-development team since 2026.
