Behind the Scenes at LeoStudio: How We Actually Make Onboarding Videos

I want to tell you about the first time I almost quit making videos.

It was a Tuesday. I was sitting at my kitchen table—the same chair I still use, the one my brain has learned to associate with work—staring at a script that a client had sent me. It was perfect. Grammatically flawless. Well-structured. Polished in every way a text can be polished.

And it was completely empty.

The client had generated it with ChatGPT. They were proud of it. They'd sent it with a note that said, "Ready to shoot." I read it three times. I couldn't find a single sentence that told me anything real about their product, their users, or the problem they were solving. It was a beautiful shell with nothing inside.

That's the moment LeoStudio changed. Not the moment the company was founded. The moment the company found its actual job. We weren't in the business of animating scripts. We were in the business of finding the story that a script was supposed to contain. And that story is almost never in the first draft—whether the draft was written by a human or by a machine.

This is the behind-the-scenes story of how we make onboarding videos now. The team. The process. The interview that changed everything. The mistakes we still make. And the philosophy that runs through every project we take—the belief that a video isn't content, it's an intervention, and an intervention only works if it meets a real moment with a real message.

How LeoStudio Started (And Why It Almost Didn't)

I started LeoStudio because I kept watching founders explain their products badly.

Not because they were bad communicators. Because they were too close to what they'd built. A founder who has spent three years inside a product can't see it the way a new user sees it. They know too much. They explain the architecture before they explain the outcome. They talk about features before they talk about the problem. They say "seamless integration" when the user is still wondering what the product is for.

I built explainer videos for those founders. Animated videos. Motion graphics. The kind of thing that makes a complex product feel simple. And for a while, that was the whole business. A client sent a script, we animated it, we delivered it, we moved on.

Then two things happened. The first was AI. Suddenly every founder had access to a tool that could generate a script in ten seconds. The scripts got faster. They didn't get better. In fact, they got worse in a specific way—they got more generic. The AI could produce the shape of a script without producing the substance. And founders, who were already too close to their products, couldn't tell the difference.

The second thing was SaaS. We started getting more software clients. And software clients had a problem that other clients didn't. Their users weren't just confused about the product. They were confused about the next step. They signed up, logged in, and froze. And no amount of beautiful animation was going to fix a script that didn't tell them what to click.

That's when the Anthropological Interview was born. Not as a clever name. As a survival mechanism. We couldn't animate the AI scripts. They were hollow. So we started asking the questions the AI couldn't ask. And the answers became the script.

You can see the results of that shift on leostudiohq.com. The case studies aren't about animation styles. They're about moments—the moment a user gets stuck, the moment a founder explains the one thing that matters, the moment the video earns the click.

The Team: Remote, Small, and Built Around the Work

I want to be honest about the team, because most agency "about" pages are fiction.

LeoStudio is remote. Not "remote-friendly." Remote. My whole team works from wherever they are, and we've all been there—the kitchen table, the coffee shop, the couch that's too comfortable to be productive. We talk about remote work distractions openly, because pretending they don't exist doesn't help anyone. The Pomodoro technique. The focus bubble. The one task written down the night before. Those aren't corporate wellness initiatives. They're the actual survival habits of a distributed team.

The team is small. Ten to fifty people, depending on how you count. But the core production team is smaller than that—a handful of people who do the work. A scriptwriter. A motion designer. An animator. A project lead. And me, doing whatever the project needs on any given day.

That smallness is a feature, not a limitation. A small team can't afford to hide behind process. A small team can't pass a project down a chain of seven people, each of whom adds a layer of polish and removes a layer of meaning. A small team has to get it right the first time, because there's no one else to fix it.

This is the part of LeoStudio that I'm proudest of. We're not a studio that scales by adding people. We're a studio that scales by getting better at the first conversation. The Anthropological Interview is how we do that. The team is built around that interview. Everything downstream is execution.

The Anthropological Interview: The Step That Changes the Video

Here's the process. It's not complicated. It's just specific.

When a client sends us a script—whether they wrote it or an AI wrote it—we don't start production. We start with a conversation. Three questions. The questions AI cannot answer, because the answers aren't in any training data. They're in the client's head, and they only come out when someone asks.

Question one: If the viewer forgets everything else, what is the one message they must remember?

This question finds the core. Not the feature. Not the benefit. The one thing. Most scripts try to say twelve things. The Anthropological Interview forces the client to pick one. And when they pick one, the script writes itself.

Question two: Tell us about your most satisfied client. What was their struggle and their "aha!" moment?

This question finds the heart. Not the product. The human. The specific person who had a specific problem and experienced a specific moment of relief. That moment is the video. It's the emotional payoff that makes the viewer keep watching.

Question three: What is the most difficult thing to explain about your offer, and how do you personally describe it?

This question finds the simplicity. Not the marketing language. The way a human would explain it to a friend. The explanation that doesn't use the words "platform" or "seamless" or "robust." The one-sentence version that a real person would actually say.

Those three answers become the script. We merge them with the structure the AI provided—the AI is a good junior writer, a terrible strategist—and we produce a hybrid. The AI's skeleton. The client's soul.

The first time we did this, the client was skeptical. They'd spent money on the AI script. They didn't want to start over. Then we showed them the hybrid script. It was shorter than the AI version. It was simpler. It opened with a problem instead of a product. And when we read it out loud, it sounded like a person. The client approved it in one round.

You can see the same thinking on leostudiohq.com. Every video starts with the moment, not the message. The Anthropological Interview is how we find the moment.

The Production Pipeline: From Brief to Delivery

Once the script is approved, the production moves fast. This is where the AI-human workflow earns its name.

Ideation and storyboard. AI generates rough storyboard concepts from the approved script. A human refines them. The AI is good at volume—it can produce twenty storyboard variations in seconds. The human is good at judgment—picking the three that actually work, and knowing which one is right.

Screen recording and asset preparation. For onboarding videos, this is the heart of the production. We record the product screen in a clean demo environment. No test data. No dev clutter. A cursor that moves deliberately, at roughly half the speed that feels natural. Because the viewer is seeing the interface for the first time. Your sense of pace is wrong by default.

Motion design and animation. This is where the AI does the heavy lifting. Color grading. Transitions. Basic motion. The human focuses on the parts that carry meaning—the win moment, the emotional beat, the pacing that keeps the viewer watching.

Voiceover. We use a human voice for the main onboarding videos. Not a voice actor performing. A person talking. The tone is conversational. The script was written for the ear, not the eye. And the voice matches the relationship—if the video is from a founder, the founder records it. If it's from the product, a warm human voice works. AI voices are fine for localization and internal content. For the main video, the voice has to sound like a person who knows how the user feels.

Editing and polish. This is where the video becomes a video. The screen recording gets cut. The voiceover gets synced. The captions get added. The pacing gets fixed. The video gets cut in half, then cut again. The first draft is always too long. The final video is always shorter than anyone expected.

Delivery. Three exports. 16:9 for desktop and web. 1:1 for social. 9:16 for mobile and email. Captions burned in for mobile. A separate caption file for desktop. Compressed for web. Tested on a phone before delivery.

The whole pipeline takes three to five weeks for a 90-second onboarding video, assuming the client makes decisions quickly. The fastest project we've ever done took nineteen days. The slowest took eleven weeks. The difference was never the production. It was the decisions—the feedback rounds, the stakeholder reviews, the "can we just add one more feature" requests that arrive in week four.

What We've Learned the Hard Way

I want to be honest about the mistakes, because the failures are more instructive than the process.

We used to animate scripts without questioning them. The AI script fiasco taught us that a script can be well-written and still be wrong. Now we interview before we produce. Every time. No exceptions.

We used to open videos with the brand. "Welcome to [Product]." Users dropped off in the first fifteen seconds. Now we open with the user's pain. "You signed up to get answers. But right now, you're staring at an empty screen." That line works. The brand opener never did.

We used to design for desktop. We tested on our monitors. We forgot that more than half of users sign up on a phone. Now we design for mobile first and scale up to desktop. The mobile constraint forces clarity.

We used to ignore the mute viewer. Most people watch onboarding videos with the sound off. Our first videos depended on the voiceover to carry the meaning. Now every key step has on-screen text. The video works in silence.

We used to forget the landing page. We delivered the video and assumed the client would handle the rest. They didn't. The video sat on a page with no autoplay, no captions, no CTA. Now we build the landing page as part of the delivery. The video and the page are one asset.

We used to treat the video as the deliverable. It's not. The video is an intervention. It has a trigger, a placement, and a metric. If any of those are wrong, the video fails—no matter how good it is.

The Philosophy: Clarity Beats Features

I want to end with the thing that actually matters, because the process only works if the philosophy is right.

Most SaaS videos fail because they try to show everything. The product has twelve features. The video shows twelve features. The user watches twelve features and remembers none of them.

We build videos that show almost nothing. One action. One win. One next click. The video ignores 80% of the product. That feels terrifying to the client. It also works.

The reason is simple. A user who is stuck doesn't need a feature tour. They need the next click. A user who is confused doesn't need to understand the platform. They need to understand what to do next. A user who is about to leave doesn't need to be impressed. They need to be rescued.

That's the whole philosophy. The video is a rescue, not a lesson. It meets the user at the moment of maximum confusion and shows them the one thing that matters. Everything else can wait.

This is what the Anthropological Interview is for. It finds the one thing. This is what the AI-human workflow is for. It produces the one thing faster. This is what the team is for. It protects the one thing from the twelve features that want to crowd it out.

You can see the results on leostudiohq.com. The case studies aren't about animation. They're about moments. The moment a user gets stuck. The moment a founder explains the one thing. The moment the video earns the click.

The Final Word

LeoStudio isn't a video production company. Or at least, that's not how I think about it.

We're a clarity company. We take complex products and find the one thing that matters. We take AI-generated scripts and inject the human story that makes them worth watching. We take founders who are too close to their products and translate what they know into something a new user can actually understand.

The videos we make are 90 seconds long. They show one action. They open with the user's pain. They work on mute. They work on mobile. They end with permission to ignore everything else. And when they work—when they really work—they don't just get watched. They change behavior. They reduce time-to-value. They reduce churn. They make the product easier to use, not because the product changed, but because the user finally understood what to do.

That's the job. Not animation. Not production. Clarity.

If you're building a SaaS product and you're trying to explain it, go to leostudiohq.com. Look at the case studies. Read the thinking. Then ask yourself the three questions we ask every client. If you can answer them, you don't need us. If you can't, that's the work.