EdTech
Lingoverso
In-house product · EdTech
Problem
After more than a decade teaching languages — English and Spanish, across regular school and language courses — one thing kept repeating: the method is good, but the follow-up between classes isn't there. The student leaves motivated and, by the following week, has already lost momentum, because no one is holding them accountable for practice on the days without class. That observation turned into a decision: instead of continuing to just teach for other schools, build a method of my own that solved that gap — not with more content, but with active follow-up between classes. The technical challenge was doing this without being a developer — the answer wasn't hiring a team, it was assembling the system in layers, each one in the right tool.
Process and decisions
The end-to-end validated journey: visitor → website (lingoverso.com) → level quiz (Tally, 14 questions) → estimated level automatically saved to the CRM (Make → Brevo) → free trial lesson scheduled (Calendly) → student enrolled → reinforcement practice with the pedagogical chatbot "Lingo," which corrects by CEFR level. This is the part of the system that already runs without manual intervention at every step.
A brand identity of its own, documented: primary green color (#568985), serif typography in headings (Lora) to reinforce "serious method, not a generic course" — never generic, never infantilized, that's the literal guideline in the brand guide. On pricing, couples or friends pay the same price per person: the difference between individual and paired lessons is pedagogical (class dynamic), not financial — a deliberate decision not to use price as a differentiator where the real differentiator is the lesson experience. Payment happens via WhatsApp and a link, with no integrated checkout: prioritizing real impact over complexity at this stage, an MVP that works now, without premature over-engineering. And the chatbot corrects at most once per response — correcting everything at once overwhelms the student, correcting little and consistently sustains the habit.
Result
Live since 2025. The acquisition funnel (quiz → CRM → trial lesson) is validated end to end and runs without manual intervention at every step. The flashcards app, rebuilt in Next.js + Supabase, is also ready — it just needs testing and optimization for the offline version. The WhatsApp support automation via Telegram (n8n) is also ready, and like the rest of the system, keeps getting refined. The strongest signal so far is qualitative: acceptance of the practice chatbot and the flashcards app among students who have already tried them was well above expectations.
Roadmap
International site (Webflow), redirected to teaching Portuguese to foreigners.
Building something similar?
If you're building an EdTech or educational product with AI, let's talk.