Led by Park Min-jun, Wrtn AX leverages chatbot experience to build autonomous AI systems for public institutions and enterprises, targeting 10 billion won sales by 2025
Generative artificial intelligence (AI) startup Wrtn Technologies (Wrtn) has entered the AX (AI transformation) business. On the 29th of last month, it established an internal independent company (CIC) called ‘Wrtn AX.’ Park Min-jun, 30 years old, who previously served as head of planning and strategy at Wrtn, will lead Wrtn AX.
AX refers to the integration of AI technology into business systems or workflows. The core of the AX business is the ‘AI agent.’ An AI agent is an autonomous AI system that makes independent decisions when performing specific tasks. Wrtn AX is currently pursuing a business model that builds customized agentic AI for public institutions and enterprises.
Park is one of the six original members who founded Wrtn alongside CEO Lee Se-young. He graduated from Seoul National University’s Department of Physics and also studied computer engineering and financial economics. During his university years, the team preparing for a youth academic conference incurred nearly 100 million Korean won in debt after the event was canceled due to COVID-19. At the time, the team members’ average age was 25. To repay the debt, the team members worked as private tutors and online instructors. After transitioning the conference online, they achieved success and used the remaining funds to start a business.
Wrtn, which began with six members in 2021, has now grown to approximately 110 employees. Its cumulative investment reaches 130 billion Korean won. Recently, it gained recognition through an advertisement featuring model G-Dragon. Within four years of its founding, it secured 6 million monthly active users (MAU) and is now targeting 10 million.
Park worked as a data analyst at global hedge fund WorldQuant, developing AI algorithms, immediately after founding the company. After returning to Wrtn, he was responsible for the company’s new business strategies. The chatbot service and the company’s first agentic AI service were also his creations.
The reason Wrtn, a startup, established a CIC is because the AX business operates in the B2B (business-to-business) sector. Wrtn grew primarily through B2C (business-to-consumer) operations. However, the AX business, which targets enterprises and public institutions, requires independent decision-making structures due to varying AI adoption speeds and customized demands across clients.
On the 30th of last month, Park, met at Wrtn’s office near Gangnam Station in Seocho-gu, Seoul, explained, “The business DNA of B2C and B2B is fundamentally different. Given the fast-paced nature of the industry, we judged that operating independently from the existing Wrtn in terms of internal culture and decision-making processes would be more efficient.”
Park outlined four stages of AX. The first involves introducing a basic conversational interface (chatbot) to create an environment where employees can interact with AI. The second stage enhances responses by adding information to existing data.
The third stage utilizes tool-based agents. AI executes code, accesses internal systems, and controls computers. The final stage is the ‘human-like agent,’ where AI learns and adapts like humans.
Currently, Wrtn AX is building a system to implement the fourth stage. Park stated, “We have already created many basic chatbots through Wrtn’s services, and we have experience building systems for internal and external use at the second stage. We are currently operating a separate team preparing for the third stage.” He added, “We are also developing an internal system for the fourth stage.” Below is a Q&A with Park.
— Wrtn’s founding story is intriguing.
“In 2019, the team members who had worked together running an academic conference secretariat founded Wrtn. The structure involved prepaying venue fees and covering costs with participant fees. However, COVID-19 struck. A student from Wuhan, China, was scheduled to attend, and although we blocked their entry at the airport, we couldn’t stop those transiting through China. Ultimately, the conference was canceled, and we couldn’t collect participant fees. We incurred nearly 100 million Korean won in debt. We repaid it by tutoring via Zoom. After transitioning the next conference online, we successfully hosted it, repaid all debts, and had funds left for founding the company.”
— What is Wrtn AX’s competitive edge?
“Three factors. First, Wrtn has handled millions of traffic instances through its B2C chatbot services. Experiencing diverse user questions and scenarios, we understand where agents fail and how to improve them. Second is UX (user experience) capability. We have been obsessed with creating AI that is user-friendly. Lastly, consulting expertise. We don’t just provide models; we analyze a company’s workflow, diagnose bottlenecks, and build customized agents.”
— Why emphasize AX consulting?
“AX is costly due to initial expenses like graphics processing units (GPUs) and labor. Bottlenecks vary by company. Without identifying the right AI application points for their workflows, the ROI (return on investment) won’t materialize.”
— Are results emerging?
“We are working on projects with the Ministry of Culture, Sports and Tourism, Gyeonggi Provincial Office of Education, Amorepacific, and others. Requests to optimize document tasks, such as creating marketing phrases and press releases, are increasing. We are building references centered on public institutions and large enterprises, with consulting requests already reaching double digits. Currently, we are hiring due to overwhelming demand beyond our current capacity.”
— What government support does a growing company like yours need?
“GPU support is critical. Large-scale computing resources are essential for AX, but GPUs cost hundreds of millions of Korean won, creating a burden. A dedicated GPU pool for AX would be ideal.”
— What are your future plans?
“Our immediate goal is to achieve 10 billion Korean won in sales by next year. We are preparing paid AX lecture programs for executives. We are also developing educational AX services, such as an ‘AI tutor’ for students. Long-term, we aim to commercialize ‘human-like agents’ that learn and adapt like humans, deploying them to enterprises and institutions nationwide by 2030. We expect to reach a technological completion stage as early as 2027.”
