[For more diverse corporate information on the startup featured in this article, please visit the Unicorn Factory big data platform “Data Lab.”]
“Many people believe they need to study English, but after about two weeks, the frequency with which they open the app drops sharply. Because studying English feels like doing homework, even opening the app becomes a burden. So we took the opposite approach. Instead of users opening the app to study, AI (artificial intelligence) initiates a call first, allowing for natural conversation.”
Park Hyun-mu, CEO of Hypernova, introduced his AI-based phone English service, “HeyRing,” in this manner. Founded in October 2022, Hypernova explored the market by consecutively launching B2C services from its early stages.
After repeating the process of launching and shutting down as many as 13 B2C services over the first two years, the company focused on the “phone English” market last year to release HeyRing. Since its launch, HeyRing has recorded a top-tier revenue ranking in the education category of domestic app stores.
Park said, “After making the strategic decision to focus on what can impact the market immediately rather than what AI will change in the future, we examined the education market within the B2C sector.”
Subsequently, through user interviews conducted over three months, they confirmed demand for phone English. They also considered that phone English is a well-developed learning method domestically. It was explained that video English has established itself as the primary format overseas.
“Like a Friend Rather Than a Lesson”…The Conversational Approach Chosen by AI Phone English
HeyRing structured its service in the form of everyday conversations with a friend, rather than taking classes at set times like traditional phone English services. Park said, “The fundamental reason people want to learn a language is ultimately the desire to communicate with someone, but existing services failed to address this aspect,” adding, “We designed it to provide customized learning content based on call recordings after conversations end.”
HeyRing offers a paraphrasing feature that displays users’ utterances converted into natural expressions used by native speakers during conversations. After a call ends, the service also provides a report summarizing fluency, speaking speed, number of words used, and key expressions.
Users can ask additional questions based on the report content. Park noted, “While there are no right answers in language, many domestic users want to verify whether their expressions are natural,” adding, “We included a Q&A feature to resolve such curiosities.”
Another feature highlighted by HeyRing is personalizing the learning experience based on individual user conversation data. Utilizing large language models (LLMs), the service accumulates data on what topics users discussed and what expressions they used, reflecting this in subsequent conversations and learning content.
Park said, “As users continue to converse, data accumulates, enabling us to provide more personalized conversations and content based on that data.”
The correction method also varies by user. Park explained, “Users who need immediate confidence to start speaking English have different needs than those who want to refine every sentence accurately,” adding, “The AI assesses this and adjusts the intensity of corrections according to the user.”
“Competition is Fierce, but No One Has Kept Users Engaged Long-Term”
The English education market is a sector where various services compete. Park viewed the market’s competitive landscape with a focus on how consistently users utilize the service. He said, “From a marketing perspective, competition is fierce, but from a retention perspective, it is not,” adding, “While there are many English services domestically, we believe no company has established a definitive retention moat.”
Park emphasized that the experience of AI initiating calls and the intimacy formed through ongoing conversations are HeyRing’s differentiating factors. He said, “We determined that we can create our own competitiveness through the experience of receiving calls and the relationship built with AI, and grow based on this.”
Hypernova secured seed investment at Spring Camp in July 2023. Subsequently, selected for TIPS (Technology Innovation Promotion System), it is researching and developing multimodal-based personalized conversation technology.
Currently, the company is continuing its business based on its own revenue rather than additional investment. Park stated, “We received investment only once, and have been operating solely on operating profit since then,” adding, “Rising to a top-tier revenue ranking in domestic app stores with a small team of about six people without spending heavily on marketing is a result confirming that there is actual demand in the market.”
Plans for additional fundraising will be judged based on revenue. Park said, “If we record annual revenue of 10 billion won, I think we would consider investment,” adding, “If we receive investment in a market where actual demand has been confirmed but fail to generate revenue, I believe we have not met the basic conditions for securing follow-on investment.”
Overseas expansion is also underway. Hypernova has currently launched a Japanese service and operates its service in Taiwan. It is also preparing to enter the U.S. and South American markets. Park noted, “In Taiwan, reactions were better when we explained it as a ‘call experience like with a friend’ rather than leading with the concept of ‘phone English’,” adding, “Which element to emphasize depends on national sentiments and must be adjusted to fit local conditions.”
As a service that allows conversations with AI similar to real people, there are concerns about potential misuse for purposes other than learning. Park said, “We have an internal policy of permanently suspending accounts without prior warning if they are used in unethical ways, such as sexual conversations,” adding, “We treat users distinctly from character chatting or AI companion services.” He further added, “Since many service users aim for learning purposes, conversion rates appear high in marketing that emphasizes learning value.”
Park is aligning HeyRing’s future direction with “communication” rather than “learning.” He said, “Our approach is that language should be learned through communication, not study,” adding, “Proving that this method is correct is the goal we aim to reach.”
Park first experienced entrepreneurship during his second year of university. Although he majored in life sciences, he pivoted to entrepreneurship because dedicating over 10 years to research felt too long at the time.
He explained that through continued entrepreneurial experiences, he realized he is a person who values a sense of achievement. Regarding HeyRing, he called it “my first work to prove myself” and stated he is currently focused on improving the service’s quality.
[MoneyToday Startup Media Platform Unicorn Factory]
Source link
