Content creation
How I Create Motion Design Videos for My SaaS Products with Claude
Having made motion design videos in After Effects, I wanted to try Claude Opus for my SaaS marketing. Here’s the step-by-step process I use, and where I keep creative control.
When I want to get the word out about a SaaS product, motion design is one way to show what it does rather than just describe it. I’ve made a number of videos in After Effects before, so I know how much work can go into them.
I wanted to see what a different approach might look like with Claude Opus. But to get something that genuinely matched what I had in mind, one prompt wasn’t enough for me. So I broke the production process into six stages, with a decision or approval between each one.
The goal isn’t to let AI make every decision. I use it to help with analysis, creative exploration and production, while keeping control of the message, visual direction and final result.
A video illustration
Start with the product, not the animations
Before thinking about transitions or colors, I give Claude enough context to understand the SaaS product: what it does, which problem it solves and who it is for. When those materials are available, I can draw on the source code, documentation, landing page and existing marketing content.
The aim is to avoid making something visually polished but vague or disconnected from the product. I look for one central message, rather than a tour of every feature. Then I ask for several genuinely different advertising concepts and choose the one that feels most relevant.
An animation can use visual metaphors to explain an idea. It still shouldn’t imply that a feature exists if the product doesn’t actually offer it.
I’ve documented this starting point and the expected deliverables in my six-step working notes.
Explore different visual directions
Once I’ve chosen a concept, I ask Claude to explore different visual styles. Instead of comparing still images alone, I prefer to see short animated tests lasting a few seconds. They give me a better sense of motion, transitions, colors and the overall feel.
Ideally, each test uses the same short sequence. That way I’m comparing visual directions, not different stories. I can pick one approach and refine it before moving on.
Five seconds can tell you a lot about a style and how individual movements feel. It cannot tell you whether the entire ad is paced well. That comes later, with the full preview.
Build the script and storyboard
At this point, I know what the video needs to say and what it should look like. I can ask for a complete script and storyboard: the sequence of scenes, on-screen copy, transitions and what happens visually.
I also plan space for the voiceover, pauses and sound cues. I’m not producing the final soundtrack yet. I’m checking that the whole idea will fit the intended duration. In a video of around 30 seconds, a few extra words or one long transition can throw the whole thing off balance.
The storyboard helps me fix ideas before committing to the full animation. Changing a planned scene is much easier than rebuilding it after a complete render.
Render a full, low-resolution preview
Once the storyboard is approved, I have the full video produced at preview resolution. I’m not aiming for the sharpest possible final export yet. I want to see the entire film from start to finish.
Now I can judge the overall pacing, how the animations connect, how long text stays on screen and where the narration will fit. That’s the difference from a five-second test: this is where I check that the whole film works.
If one section doesn’t work, I ask for a targeted change. There may be several rounds of refinement. The point is to settle the visuals before starting a heavier final render.
Add music, sound effects and voiceover
Once the visuals are stable enough, I move on to audio: music, sound effects and narration. I can record my own voice, work with someone else or experiment with voice processing using a tool such as ElevenLabs.
What matters is that the voice comes in at the right moment and stays intelligible without getting buried under the music. I also prefer to keep the audio tracks separate while working. If I replace a voice take or adjust its volume, I shouldn’t have to render the animations again.
At this stage, I can review something close to the intended audiovisual result without committing to the cost of a high-quality final video render.
Render the final video and check it
Once picture and sound are approved, I run the high-quality render and combine the video with the final audio mix. My production notes use Full HD at 60 frames per second as a default target. Those are working settings, not requirements for every project.
Then I check that transitions are smooth, text and logos are readable, colors look right and the audio is in sync. I also check what it looks like on a smaller screen: a composition that works on a desktop can be hard to read on a phone.
Depending on where the video will be used, I prepare a version with sound, a muted version for autoplay or a smaller web-friendly export. The goal isn’t to create as many files as possible. It’s to make the versions I’ll actually need for distribution.
What I’ve learned from this approach
This workflow doesn’t make motion design automatic. I still need to write clear instructions, review the output, make decisions and sometimes revisit earlier choices. I also don’t have a reliable number for how much time it saves: that depends on the quality bar, the complexity of the film and the number of iterations.
For me, the value is in not trying to solve concept, animation, voiceover and export in a single request. I approve one thing at a time and, as much as possible, avoid redoing work that already looks right.