Among the professions at risk of being replaced by AI, language teaching is definitely at the top of the list.
It's not necessarily because it's a good idea. Some employers, including his most recent Duolingo, have decided that AI makes sense enough to replace human experts when it comes to language instruction. Despite the fact that text translated by AI tends to have a less rich vocabulary than human translation, the cost savings are enough to make the trade-off worth it in the minds of some administrators. Attractive enough.
But some companies argue that AI can do things at scale that language teachers cannot.
One of them, Loora, uses conversational AI to teach English to students. Founded by Roy Mor and Yonti Levin, Loora's iOS app lets users chat with a chatbot that provides feedback on their understanding of English.
“Lula's Idea [came from] It's our frustration with language learning,” Mol told TechCrunch in an email interview. “Language learning apps are only aimed at beginners and casual learners. Human tutors are very expensive, inconvenient, and have limited availability.”
Loora, which means “language” in Arabic, allows learners to choose from several AI-generated conversation topics and scenarios, ranging from sports, technology, business, fashion, books, and TV shows to interviews and presentations. provided to you. The app provides feedback on grammar, pronunciation, accent, and provides direct translation in your native language if users get stuck.
Image credits: Lula
Loora scores your proficiency over time and uses this score to customize conversations to your level of conversation.
A significant number of English learning platforms offer features along these lines, including OpenAI-powered Speak, Preply (which recently enhanced its AI technology), and ELSA. But Mol insists that it differs from Loora in that it is aimed at “serious learners” who are trying to become fluent in English for personal purposes. and Professional advancement.
“Most other language learning apps on the market are limited and gamified,” Mo said. “Loora has built, trained, and optimized our AI for the sole purpose of helping users acquire English fluency that goes far beyond everyday conversational skills. We use only proprietary data and a bespoke training and evaluation system to train and optimize our models, resulting in continuous improvements in retention.”
Mor further argues that Loora is better suited for certain language learning use cases (such as pitching ideas in a business meeting) than other apps or tutors. He argues that tutors are limited by their domain knowledge, a limitation that his Loora app does not have (Mor claims). And specialized tutors are likely to be in higher demand than general, all-purpose tutors, Mol added.
“Say a learner wants to learn how to discuss business concepts at a high level for work purposes,” Mo says. “Even if the instructor is a native speaker, if he or she is unfamiliar with it, he or she may not be suitable to teach English for that specific purpose.”
This is very promising given the inherent limitations of language teaching apps, especially those without an element of human feedback.
When Michigan State University studied the effectiveness of a popular language learning app, nearly all participants improved their grammar and vocabulary, but they failed to improve oral proficiency, a common drawback of digital language learning programs. It was only about 60%. The authors of this study concluded that the best approach to learning and retaining second language skills is a hybrid setting that combines online and classroom learning.
Loora assigns a score according to the user's perceived English proficiency. Image credits: Lula
However, this did not deter Loora's investors. They may be persuaded by the size of the overall addressable English learning market (topping $70 billion by 2030, according to data analytics firm Research & Markets).
Loora today announced that it has raised $12 million in a Series A round led by QP Ventures with participation from Hearst Ventures, Emerge, and Two Lanterns Venture Partners, bringing Loora's total funding to $21.25 million. . Mo said the cash will be used to fund the development of Loora's Android app, “deepen” Loora's core AI technology and conversational capabilities, and expand its headcount from 14 to 25 by the end of 2024. It's planned.
Loora also plans to launch enterprise services that will further expand its current customer base of 15,000 app users. (Access to Loora's app costs $15 a month, or $120 a year.) The startup's consumer business has been steadily growing (in annual recurring revenue, it expects to grow eight times by 2023). ), Mol expects corporate customer growth to accelerate.
“Our planned B2B offering will make Loora available through employers, universities and institutions, making it increasingly accessible to those who need it most,” Mo said. “and [the Series A] With funding, efficient unit economics, a growing customer base and continued demand for English language learning solutions, we are well positioned to overcome potential headwinds and continue to grow and serve our learners. I believe that. ”