When COVID-19 was at its peak, the research lab closed, and there was nowhere to go. The founding members, who were part of the same lab at the time, participated in an artificial intelligence (AI) competition. They used the discomfort they felt while staying up all night as a business idea.
Coxwave is an AI startup that creates data analysis solutions necessary for the generative artificial intelligence (AI) era. Founded in 2021 by four master's and doctoral students from the data mining lab at Seoul National University, they delved into the inadequacies of existing data analysis systems while participating in AI competitions.
Coxwave introduced initial products like the image editing service 'Hama' and the generative image search 'Enterpix.' However, these services made it difficult for users to analyze how they utilized AI. After selling the two services, Coxwave shifted its focus to the generative AI product analysis platform 'Align' and entered the business-to-business (B2B) market.
'Align' is a solution that analyzes various hints left by users in AI services, identifies problems with the service, and suggests improvement directions. Its distinctive feature is going beyond simple keyword analysis to understand context and intent.
Kim Joo-won (31), CEO of Coxwave, met with ChosunBiz at the company's office in Yeouido, Yeongdeungpo District, on the 26th of last month, and noted, "When you tell a chatbot to 'find it again,' it means that you are dissatisfied with the results given by the chatbot and are asking for a more accurate answer. Existing services struggled to understand this subtle context. However, Align analyzes that the phrase 'find it again' indicates that the answer provided was insufficient, and it provides a service that helps find ways to improve clarity in responses."
Recently, Coxwave has made a name for itself in the domestic AI ecosystem through collaboration with Anthropic. Coxwave is utilizing Anthropic's 'Claude' model in its services. The emphasis on 'safety' and 'trustworthiness' in Align's service philosophy aligns with Anthropic's direction, which became the basis for the collaboration. Currently, Coxwave is supporting Anthropic's entry into the Korean market.
Coxwave, which raised 4.5 billion won in investments from KB Investment, Danal, and Seoul National University Technology Holdings in January last year, has achieved a cumulative investment of 5.5 billion won. The company is currently using India as its first foothold for global expansion and collaborating with local corporations.
—What is the background behind the creation of 'Coxwave'?
We gathered at the Seoul National University data mining lab in 2021 to start a business. The founding members included one doctoral student and three master's students, all around 27-28 years old at the time. When COVID-19 closed the lab, we began participating in hackathons and AI competitions. Many competitions required tuning AI models according to data provided by the organizers to check accuracy. While participating in these competitions, it became cumbersome to go through the process of data analysis and model training every time. The initial business idea was to create a solution that automatically analyzes data and enables AI to learn on its own with just data input.
—What is the main product 'Align'? How does it differ from competitors?
Align is an AI product analysis platform that helps improve service performance based on user data. People often express frustration or disappointment when using chatbots like ChatGPT and may stop using them. We focused on the negative experiences accumulating as data in generative AI services. Align extracts analysis elements like reasons for user dissatisfaction from conversation data between users and AI, and compresses them into about ten key points to deliver to client companies.
The strength of Align is 'personalization.' Among the extracted data, there are emotional data from users, and we refer to many psychological studies while using various scales for emotional classification. Users provide explicit feedback like likes or dislikes, along with hints of their preferences in various ways, such as clicking a copy button. This information is organized and provided to AI product developers and managers to help them easily understand.
—You operated 'Hama' and 'Enterpix' before Align. What was the reason for selling the two services?
Hama was an image editing service, and Enterpix was a generative image search service. Align began as a small analytical dashboard that we internally developed and used to enhance Hama and Enterpix services.
While operating the two services, we realized that generative AI products felt very different from traditional web or app experiences. In web or app settings, the buttons or input texts are predefined by the planner, limiting user freedom. In contrast, generative AI products allow more user freedom, making it difficult to predict user behavior. It wasn't easy to analyze how users input text, how they dragged to delete specific parts, or how they colored in Hama and Enterpix services. There weren't any analysis tools for generative AI products at that time, so what we developed internally became the foundation for Align.
During the operation of the two services, as many generative AI products were emerging, we faced a dilemma: should we continue focusing primarily on image services as before, or should we provide B2B services based on the difficulties we experienced? After much consideration, we chose the B2B service route and sold the two services to Sandoll when the opportunity arose.
—You are currently targeting the Indian market. What is the reason for focusing on India?
India has a large population and diverse culture, which allows Coxwave to explore many opportunities. The adoption of AI is also proactive, and various startups are emerging with relatively fewer regulations. There are many executives of Indian origin in the U.S., which can aid in expanding into the American market, and proximity to the Middle East, including Saudi Arabia, is expected to help in expanding within the Asian market. We have signed an MOU with Renan, a data platform company in India, and are also collaborating on joint projects with PwC India.
—You are also aiming for entry into the U.S. market. Is there a set timeline?
First, we need to focus on the Indian market to secure good case studies. To make a significant impact when entering the U.S. market, it seems better to achieve results first in the large global market of India. We are operating on a two-track strategy: targeting both the Indian and U.S. markets simultaneously. The transition to a joint CEO system was also to operate global expansion more efficiently.
We are consistently promoting our SaaS in the U.S. market and conducting visit meetings with U.S. corporations. The collaboration with Anthropic was part of our strategy for entering the U.S. market. We are actively participating in developer-related events held in the U.S., and we conducted advertising for Align in Times Square in New York City at the end of last year.
—If you could advise the government for the advancement of the domestic AI industry, what would you say?
To secure global competitiveness, it is essential to have opportunities to go abroad. There is a wide range of support for early startups, and we actually received a great deal of help. However, there comes a vague period after the initial stage. Particularly, AI takes a long time to develop properly, and the support tends to be concentrated in the early phases, so it would be beneficial if more diverse support was available as companies transition into a mature stage. It's not just monetary support that is needed; support that aids in overseas expansion is essential.