Speak is an AI-powered language learning platform founded in 2016 by Connor Zwick and Andrew Hsu, specifically designed to overcome the challenge of attaining spoken fluency.
Its core mission is to reinvent the way people learn by providing an accessible, human-level AI tutor that replicates the experience of a one-on-one conversation, allowing users to practice speaking without the social pressure or high cost of a live teacher.
Unlike many competitors that focus on grammar and vocabulary quizzes, Speak’s philosophy emphasizes speaking out loud as much as possible to build confidence and automaticity.
Rationale
The Edtech industry sits at the intersection of education, business, and technological evolution, and this marketplace has seen rapid innovation in recent years. Among the fastest-moving segments of this market is conversational AI, which is projected to become an approximately $80 billion global industry by 2033–2034. Recent advances in LLMs have accelerated a shift from scripted chatbot interactions toward agentic systems capable of sustaining human-like dialogue. In educational contexts, this creates the possibility of conversational learning experiences that move away from traditional, curriculum-oriented digital instruction, and toward individualized tutoring and Socratic coaching.
Language learning represents a particularly compelling industry for this evolution to develop. More than two billion people worldwide are learning a language, while access to sustained conversational practice remains limited by cost and availability of qualified instructors. Of the four main skills taught in language training (reading, writing, listening, speaking), only speaking requires a real-time partner. For many learners, especially adults, the primary barrier is therefore not access to content, but access to conversational interaction. This mismatch creates conditions under which conversational AI may generate original educational value by redefining what is educationally possible, rather than simply by digitizing existing approaches.
Speak has become notable in this growing market due to the speed of its growth, investor confidence, and targeting of the market opportunity mentioned above. Backed by major institutional investors and reaching unicorn status in 2024, the company positions itself not as a general AI or English-learning product, but as a specialized learning app built around spoken language practice. Its proprietary speech infrastructure and large-scale learner interaction data create a data flywheel, which, in combination with its presence in both consumer and enterprise markets, make it an especially relevant case for educational venture analysis.
Rather than evaluating whether the product is educationally appealing, this analysis adopts the perspective of a Educational Technology Venture Capitalist, examining whether Speak represents a credible, scalable use of capital within the evolving learning technologies marketplace.