The Power of Active Engagement: Why Teuida Chose Real People Over AI

In an era where every language learning app seems to be racing toward AI-powered conversations, Teuida made a counterintuitive bet: real people, even in pre-recorded videos, create better learning outcomes than synthetic interactions. It's a decision that's helped them reach over half a million monthly active users while most of their competitors chase the latest large language model.

On the Levels Podcast, Ji Woong from Teuida walked Jason and Charlie through the philosophy behind their product design—one that prioritizes active engagement over technological novelty. For product teams building consumer apps, their approach offers important lessons about what actually drives retention and behavior change.

Swimming Against the AI Tide

The language learning space has become an AI showcase. Platforms compete on how natural their AI tutors sound, how quickly they can adapt to user mistakes, how seamlessly they can generate personalized lessons. It's impressive technology, but as Charlie pointed out, Teuida took a completely different path.

Yeah, I think it was really interesting. It's a different approach. A lot of platforms, you know, kind of use AI or use those kinds of ways, but Ji Woong was really focused on actually having the real people behind it, which I thought was interesting.

This wasn't a rejection of technology for technology's sake. It was a deliberate choice based on how people actually learn languages. The Teuida team recognized that the core challenge in language learning isn't access to information or even personalized feedback—it's building the courage and habit of actually speaking.

Forcing the Uncomfortable Moment

Jason explained the product experience in simple terms that highlight why it works:

They're a Korean and Japanese language learning app where you talk to like a video of someone that's like pre-recorded speaking the language. So if you're trying to learn how to like order a coffee or something, then you have to actually speak to someone who's on video having the interaction with you.

The word "have to" is doing important work in that description. This isn't optional practice or supplementary speaking exercises. The product forces users into the uncomfortable moment of speaking out loud. You can't tap multiple choice answers. You can't skip ahead. You have to speak.

This design creates what gamification experts would call a "forcing function"—a mechanism that makes the desired behavior unavoidable. In Teuida's case, that behavior is verbal production, which happens to be the most effective but most avoided aspect of language learning.

The Theory Behind the Design

The pedagogical foundation of Teuida's approach is straightforward, as Jason articulated:

And the idea is that you learn language better when you're actually forced to speak it.

This isn't controversial among language educators. Comprehensible input is valuable, but production—actually forming sentences and speaking them aloud—is where fluency develops. The problem is that production is uncomfortable, especially for beginners. It's vulnerable. It's awkward. Most people will avoid it if given the option.

Traditional classroom settings solve this through social pressure and instructor requirements. Digital language learning apps struggle with the same challenge because users can simply close the app when things get uncomfortable. Teuida's video-based interactions create a middle ground: less intimidating than speaking to a live person, but more forcing than a passive listening exercise.

Why Real People Matter

With AI capable of generating realistic conversations, why stick with pre-recorded videos of real people? The answer likely lies in authenticity and emotional resonance. When you're learning to order coffee in Korean, you're not just memorizing phrases—you're preparing for a real social interaction with a real person.

Practicing with a video of an actual human creates a closer psychological simulation of that real-world moment. The facial expressions, the timing, the natural pauses—these elements are harder to replicate with AI and they matter for building confidence in real-world application.

There's also something to be said for the curated nature of pre-recorded content. Each scenario can be perfectly crafted to hit specific learning objectives. The "teacher" in the video delivers the exact right level of challenge, uses the exact right vocabulary, and creates the exact right pacing. AI might get there eventually, but precision and consistency matter when you're building learning pathways.

Active vs. Passive Engagement

What Teuida really nailed is the distinction between active and passive engagement. Most digital products optimize for passive engagement because it's frictionless. Users scroll, watch, read—all low-effort activities that keep session times high and drop-off low.

But passive engagement doesn't drive behavior change. If you want users to actually learn a language (or build any new skill), they need active practice. They need to do the thing, not just consume content about the thing.

This creates a product design challenge: how do you make active engagement compelling enough that users choose it over passive alternatives? Teuida's approach of situating speaking practice within realistic video scenarios provides context and purpose. You're not just repeating phrases into the void—you're having a conversation, even if it's pre-scripted.

The Retention Implications

Active engagement patterns tend to create stronger retention than passive ones, assuming you can get users over the initial friction hump. When users invest effort—when they actually speak out loud, record themselves, complete the interaction—they develop a stronger relationship with the product.

This connects back to how Teuida tripled their LTV within a year. Products that drive real behavior change and deliver tangible results have fundamentally different retention curves than entertainment apps. Users stick around because they're making progress, and they can feel that progress in their growing confidence to speak Korean or Japanese.

Lessons for Product Builders

Teuida's success with a deliberately non-AI approach offers a useful counterpoint to the current technological zeitgeist. The lesson isn't "don't use AI"—it's "don't use technology for its own sake." The question should always be: what drives the outcome we're trying to create for users?

For Teuida, that outcome is language fluency, which requires speaking practice. Pre-recorded videos of real people created the right balance of structure, authenticity, and forcing function to drive that behavior. Your product's optimal solution might look completely different, but the principle remains: design for the behavior change you need, not the technology that's trendy.

Key Points

• Teuida deliberately chose pre-recorded videos of real people over AI-powered conversations, prioritizing authentic learning experiences
• The product forces users to actively speak rather than passively consume content, creating a "forcing function" for the most effective but most avoided learning behavior
• Active engagement drives stronger behavior change and retention compared to passive consumption, even though it creates more initial friction
• Real human interactions, even pre-recorded, provide authenticity and emotional resonance that supports confidence building
• Technology choices should serve the desired user outcome rather than being driven by what's currently trendy or impressive

Listen to the full conversation with Ji Woong on the Levels Podcast to hear more about Teuida's product strategy and how they built their Instagram community.