The app is just a container

I kept quizzing my kid over text.

Random questions while she was waiting for the bus, tram etc. What's the capital of Portugal. What's a mitochondria. Spell necessary.

It took me embarrassingly long to notice that the thing my own app was asking everyone to download for was something I was already doing in a thread.

I'd been thinking about distribution for months at that point. Lexie has good retention. People who find it tend to stay and pay. But finding it is the problem. The app store is a graveyard. TikTok reach from Finland has a ceiling. SEO takes a year to compound. Every channel I tried felt like dragging people uphill toward an app icon.

And there I was, doing the actual product, in a text thread, for free, because that's where my kid already was.

So I built Lexie for WhatsApp. You text her a photo of your notes, she sends back a study set link that opens in your browser. You can also ask her to quiz you in the chat, or explain something. Flashcards, quizzes, practice exam questions, all the same stuff as the app. Just delivered through a thread you already have open instead of an icon you have to install. The part I didn't fully understand until I shipped it: this isn't a smaller version of the app. It's a different shape entirely.

WhatsApp is good at conversation and bad at flashcards. The browser is the opposite. Trying to do both in one place is always going to be worse than doing each in the place that's already designed for it. So the chat is where you ask, where you get nudged, where Lexie remembers what you've been working on. The browser is where you actually study, where the cards flip and the quizzes run and the spaced repetition does its thing. Each medium does what it's good at.

The chat-as-interface side surprised me more than I expected. When I tested with a 15-year-old, she didn't mind seeing Lexie's messages later. She didn't mark them as read and forget them. There was no sus energy about it, no corporate AI vibes, no productivity-tool stink. It just felt like another thread. Teenagers have built up extraordinary antibodies to anything that smells like an ad, a productivity tool, or a parent's idea of fun. Study app notifications fall into that category instantly and get muted within a week. But they don't have antibodies to texting. Texting is just where they live. It's the medium they trust by default.

So the question stopped being "how do I get teenagers to use a study app" and started being "what's the smallest thing I can put in the channel they're already in."

Which leads to the bigger thing I've been circling for a while. Lexie has never really been an app to me. It's a system that encodes how people actually learn. Retrieval, spacing, productive struggle, feedback at the right moments, the boring stuff that actually works. The iOS app is one container for that system. WhatsApp is another. The web app is a third. There will be more.

The part most founders get wrong, I think, is treating the first container as the product. You build an app, it works, you start optimising the app, you fight for app store rankings, you A/B test the onboarding, you tune the paywall. All of that is real work, and none of it is the product. The product is the system underneath. The container is just where the system happens to live this week.

Once you see it that way, distribution stops being a separate problem you solve after building. It becomes the same problem as building. Each new container is a new way for the system to meet someone where they already are. Some people are already in WhatsApp. Some are on the web typing a question into ChatGPT. Some are in their browser googling "how to study for AP bio." Those are all valid surfaces for the system to live on. The only thing that changes is the local shape.

That's also why I'm not precious about the iOS app being the "real" Lexie. The real Lexie is whatever feels right for the channel. On iOS that's a focused study session with image occlusion and audio review. On WhatsApp it's a thread you can text. On the web it's a guide page that turns into a quiz. Same system, different containers.

I don't have this fully figured out. Distribution is still the unsolved problem. But the WhatsApp build was the first time it really clicked that the answer isn't "find a better way to get people to my container." The answer is "go put the system inside theirs."

Text Lexie at the number on lexielearn.com/whatsapp. 5 free study sets to try it. After that, 59€ for a year, no auto-renewal.