LangListen (working name)

A research-first pitch for Aleksandra

This document is a proposal that we collaborate on developing an AI-powered language learning tool.

Personal motivations: How did I arrive here?

Language learning is a passion of mine. I have found that most language materials assume someone is learning within an academic or for career advancement reasons. Finding ways to persist and advance in language learning outside of these settings remains a major unaddressed pain point in the language learning market. For the past few years I’ve been testing out these AI-powered tools, and dog-fooding AI-based workflows for language learning as I learn my third language - Portuguese. I’ve found two successfuly strategies that inspire some of what is presented in this proposal.

One strategy is to integrate language learning into existing media consumption habits. For example, if someone enjoys romance reality shows and is learning French, they can watch “Love is Blind: France”. If they enjoy true crime podcasts they can listen to such podcasts in French. If they enjoy music they can use Spotify to learn to sing the lyrics of a French musician. If they love Harry Potter books they can read the books in French.

Several apps provide language learning tooling around media consumption habits. For example, LanguageReactor provides tooling around streaming services. Pimsleur provides audio lessons that one can listen to on their commute or while doing housework, as one might do with a podcast or audiobook.

There are media consumption habits that are not addressed by existing tools (e.g., audiobooks, TikTok feeds). Modern AI tools (LLMs, voice-to-text transcription, video description with VLMs, RAG systems, image/video generation, etc.) are providing new opportunities for building this type of tooling.

The second strategy is using paid language tutoring platforms like Italki, Prebly, and Verbling, as well as language exchange platforms like HelloTalk and Tandem. Indeed, for learners not pursuing career advancement or academic objectives, having conversations with new people on these platforms can be as much an end as it is a means to an end.

I have a strong drive to fight tech culture’s obsession with using AI to replace humans in human interactions rather than using AI to enhance those interactions. For example, I think “AI girlfriend” apps are bad, not because talking to agents is bad, but because they are based on the idea that “human girlfriends” are inferior, and that the value of a romantic partner is based on how well they can gratify you. Indeed, there are some apps that are seeking to suplant human tutors with AI tutors, such as DuoLingo Max.

In contrast, I believe there is an opportunity to support human interactions on these platforms with AI-tools. Indeed, research I conducted in India demonstrated that sellers on gig economy platforms in general, rather than being replaced by AI, actually tend to be power users. You’ll see that trend reflected in the tutor interview insights presented in this proposal, where many tutors use ChatGPT to generate content for use with their students. Italki is starting to integrate AI tools, but as the research in this proposal show, they haven’t yet found traction – which I believe is an opportunity.

A prototype app

I’ve built a prototype web app that includes some basic workflows for creating personalized language-learning materials (audio lessons, conversations, quizzes, reports) from user-provided content. In building the app, I’ve just focused on getting all the pieces of the tech stack to fit together. The app does not yet reflect a product vision.

Discovery research

To start building that vision, I’ve started a lean development-style customer discovery process. So far I’ve interviewed about twenty people. These have skewed mostly towards Portuguese tutors on these platforms, but have also included independent online language instructors (tutors who market their own online courses to students directly) This document summarizes findings and insights as well as the hypotheses I’ve built from these interviews.

Based on the interviews I’ve conducted so far, this document records:

• Interview analyses and insights
• My current product development hypotheses (which need validation)
• Three business model canvases (one primary, two alternatives)

The ask for you

If you’re interested in exploring co-founding together, my proposed next step for you (Aleksandra) is:

• Explore tutoring platforms.: Italki, Prebly, and Verbling.
• Survey competitive landscape: Pimsleur and Language Reactor especially, as well as tools that use AI such as DuoLingo Max and Italki Plus.
• Run ~10 discovery interviews in 2 weeks: Interview students or tutors (ideally more students). Use other platforms and this prototype as topics of conversation.
• Day-14 deliverable: Write an updated Business Model Canvas + a short recommendation memo: **double down / pivot / explore other collaboration directions.

Reading guide

• Start with Why us.
• Then explore the prototype app.
• Then read the analyses of the interviews
• Then consider the hypotheses - ideas for product vision that need further validation
• Finally, review the business model canvases

What I believe today (high-level)

• Beachhead: busy adult learners (not school/career-driven) who want “real progress” without high friction.
• Primary bet: audio-first “personalized podcast feed” experience at roughly $12/mo.
• Distribution wedge: tutors as a free acquisition channel (inviting learners), while learners are the payer.
• Backbone: cross-lesson memory + progress visibility (continuity and “wins”).

These are hypotheses, not conclusions. The goal is to test them fast and honestly.