How we built a smarter academy 10x faster

We take pride in having built a feature-rich platform that we continuously improve. We keep updating existing features and releasing new ones based on the feedback and comments that we receive from our customers.

The thing is, as we update our platform, we need to update our customer training material as well.

Sounds familiar? If you’re in the SaaS industry, you probably know what this process entails and how cumbersome it can be.

In our recent webinar, we discussed how we turned our knowledge base into an on-demand, video-based training academy, without a big production team or additional budget. We wanted to show you how we did it, so you can do it too.

You can jump directly to the webinar or get the gist right here from this blog post.

Part 1: Setting the stage

Why did we build a training academy in the first place? Let’s see the challenges that prompted us to offer customer training and why, ultimately, we decided to change the way we build content.

Why did we even build an academy?

Well, for one, because we sell academies. So we should practice what we preach. But we had other reasons too. We wanted to:

• Scale our product training with on-demand support resources
• Lower support ticket volume (this, in turn, would reduce our support agents’ workload, improve the quality and speed of customer service, and reduce customer support costs)
• Help customers realize our product’s full potential and improve adoption
• Increase customer satisfaction and retention
• Promote upsells

What was wrong with our academy

As much as we’d like to think that we had achieved perfection, we were not quite there yet. Why?

• Training videos were outdated and uneditable
• Only one person knew how to produce videos
• As a result, it was hard to keep our academy up to date, let alone scale

Every update meant one person redoing everything.

We needed to change that. We need to go from “Only Eliza knows how to do it” to “Anyone can do it, with systems and AI.”

We decided to rebuild the academy with a team that consisted of 3 Learning Designers, 1 AI Specialist, and the help of our customer support agents.

Part 2: Designing the syllabus

Designing a syllabus before jumping to course creation is critical.

We used three different sources of truth to guide our content creation: our competitors’ academies, our knowledge base, and our incoming support tickets and product analytics.

Source A: Our competitors’ academies

We used AI to analyze what courses our competitors are offering, which topics they were tackling, and their tone of voice. This would give us an idea of what matters the most to users of products similar to ours.

Source B: Our knowledge base

We have more than 800 articles on our knowledge base, built organically from users’ pain points and usage patterns. We used AI (again) to identify and map out clusters and translate them into course titles.

Source C: Product analytics and support tickets

We fed AI with a large set of data from our product analytics and support tickets to identify feature drop-offs, common questions, and points where customers get confused during onboarding. We turned support signals into curriculum priorities.

In just one day, we compiled the proposals from these three sources and asked ChatGPT to provide us with 3 different track proposals:

• Featured-based
• Goal-based
• Customer journey

We decided to go with the goal-based approach. This means that we built our curriculum mainly focused on what the user wants to achieve: create their first course, launch their academy, measure learner progress, etc.

ChatGPT generated not only course suggestions but also a suggested outline for each course.

Part 3: Where we started from

We already had some raw material that could be modified and repurposed into training videos. Reading this, you’ll realize that you have too:

• Support articles from our knowledge base
• Blog posts, PDFs, and FAQs
• Customer support workflows

Articles worked great as a starting point for us because they’re accurate and well-structured, and they’re based on real user questions to our customer support team.

But, they’re not plug-and-play scripts.

Convert an article into a video as it is, and it will be robotic, dry, and overwhelming with too many details. Tone, pacing, and visuals need to be reworked.

A good video script should have:

• learner-centric intro
• pauses and emphasis where appropriate
• visual storytelling cues

Part 4: From article to script

What does it take to transform a less-than-exciting support article into an engaging video script?

Step 1: Analyze the article

Some questions we asked:

• Who is it for?
• What’s the starting point and the end?
• What kind of article is it?

This helped us identify key actions and flows, and UI elements to show.

Step 2: Prompt AI to draft a script

We created a custom ChatGPT prompt so that we don’t have to repeat this process every time we want to ask something similar. The example above is a simplified version of the actual prompt we created.

Step 3: Refine the script

Next, we reviewed AI’s output and fixed some flaws, like:

• Missing steps
• Filler text
• Robotic tone and phrasing
• Wrong sequence

We also added some visual/audio alignment markers like pause, zoom, show, and click.

Part 5: Building the audio

Both audio and visuals are equally important for an engaging video that successfully transmits knowledge. Using an AI narrator for your videos is preferable because it is so much easier to update your content later on without being dependent on a single person.

As you build the audio for your video, make sure to get these two things right:

• Clear voice for better engagement
• Good timing/sync for better comprehension

We followed this simple 5-step workflow to add audio:

1. Finalize the script
2. Add pauses, chunk steps
3. Generate audio using ElevenLabs
4. Playback and sync during screen recording using Camtasia
5. Edit and trim on Camtasia

Part 6: Screencasting and visual rhythm

What we wanted to achieve with screencasting:

• Match the narration
• Keep it clean, focused, and calm
• Make every click count
• Make it straightforward and easy to recreate

Here’s what to pay attention to:

• Ensure zero distractions: close all other tabs and turn off notifications
• Zoom set at 100%
• One clip = one clear user step
• Record in silence, in 10–30s chunks
• Name clips immediately

Also, here are some tips to ensure the viewer can easily keep up with you:

• Hover before clicking
• Make slow, deliberate movements
• Pause on key UI moments
• Show one step per clip

Part 7: Compile everything

In this last part, we put everything together.

Step 1: The assembly process

1. Import assets into Camtasia
2. Align tracks:
   • Voice (top)
   • Screen (middle)
   • Music (bottom, soft)
3. Trim & tighten timing
4. Match cursor to narration
5. Freeze frames / slow-mo if needed
6. Add intro/outro templates

Step 2: Polishing and transitioning

Some points to pay attention to here:

• Fade in/out at start & end
• Blur sensitive data, like school/login data
• Add zooms or focus areas
• Use minimal callouts

Step 3: Export and upload

Our export checklist:

✅ Tight sync
✅ Cursor is clear
✅ Music balanced (not too loud)
✅ Transcript generated
✅ Upload to LearnWorlds + subtitles

The verdict: Was it faster?

We produced two videos—“Create Your First Course” and “Customize the Layout Page”—using the traditional video creation method and AI-assisted creation to see how they would compare. Eliza, who is an experienced creator, created her video without using AI. Xenia, who had never created a video before, used AI.

To sum up, AI scripts

• only perform well with structured prompts + custom GPT
• often misalign with UI steps
• require deletion, rewriting, and reordering

But they also…

open the door to non-creators
allow colleagues to contribute scripts or rough cuts

So, ultimately, they help speed up the video creation process and therefore the creation of your academy.

Results in numbers

• 3 people, 20 maintainable videos per day
• 1 small course in 1 day
• 5 courses in less than 10 days

Using AI worked for us, and we can confidently say that it will work for you too.

It’s natural to feel intimidated at first, especially if you have no previous experience with video creation or AI tools.

But as Xenia put it, the first successful video was enough to make her feel more confident and say, “We can do this!” Start with experimentation in mind instead of perfection, and the more you keep playing around, the better you’ll get.

In the meantime, watch the full webinar to dive into more details about the process we followed and find inspiration to kickstart your training academy.

And if you want to see firsthand how you can build a scalable training academy, try LearnWorlds now with a 30-day free trial. You’ll be amazed by our platform’s rich learning features and business tools.

Androniki Koumadoraki

Androniki is a Content Writer at LearnWorlds sharing Instructional Design and marketing tips. With solid experience in B2B writing and technical translation, she is passionate about learning and spreading knowledge. She is also an aspiring yogi, a book nerd, and a talented transponster.