[AI Revolution](171) AI Works Bets on 'AI Quality Validation'... Evolving into a Data Technology Company

From Testworks to AI Works: A New Name, a New Vision
Rebounding with a Focus on AI Reliability and Work Automation
From Data to Modeling to Agent Validation
Expanding AI Quality Validation in Telecom and Finance
Sales Expected to Reach 18 Billion Won This Year, IPO Planned for 2026

Yoon Seokwon, CEO of AI Works, is being interviewed by Asia Economy at his office in Songpa-gu, Seoul. Photo by Park Yujin

In an era where artificial intelligence (AI) has become a part of everyday life, the most important question is whether it can truly be trusted and relied upon. In a recent interview with Asia Economy, Yoon Seokwon, CEO of AI Works, stated, "We create AI that helps people," adding, "AI should not just be technology-centric, but must be able to function in real-world tasks." Formerly known as Testworks, the company changed its name this past April, declaring a shift from a data labeling company to a specialist in AI quality validation and work automation.

CEO Yoon explained, "The letter 'I' in our logo is tilted like a person," emphasizing, "This signifies our commitment to maintaining a people-centered identity." The 'WORKX' in the company name represents AI that operates in actual work (Work) and infinite possibilities (X). He stressed, "AI must now move beyond demonstration purposes and become a colleague that works alongside people in the field."

AI Works' business is built on three main pillars: collecting, processing, and inspecting unstructured data such as text, images, and audio for AI data projects; software testing and quality validation that have evolved through AI-based automation; and the development of AI agent solutions used in call centers (AICC) and document automation.

CEO Yoon particularly emphasized 'agent validation.' He said, "We have entered an era where AI agents oversee entire software systems, not just individual models," and added, "Now, we must validate not only model performance, but also intent recognition, guardrails, usability, and the minimization of hallucinations."

AI Works has secured quality validation projects for AI agents from all three major telecom companies in Korea-SK Telecom, KT, and LG Uplus-and has demonstrated its technological prowess by outperforming competitors in a proof-of-concept (PoC) project with KB Card. Yoon highlighted, "We excel in 'field validation' that solves real customer problems," and emphasized, "There are few companies in Korea that manage end-to-end (E2E) quality to prevent AI malfunctions."

The company experienced its first-ever negative growth last year, with sales reaching 9.9 billion won. This was due to a combination of reduced government research and development (R&D) budgets and an economic downturn. However, the situation has changed this year. Yoon projected, "Sales are expected to recover to around 18 billion won this year," and predicted, "An additional 30 to 40 percent growth is possible next year."

AI Works recently joined the NC AI Consortium, part of the independent AI foundation model project led by the Ministry of Science and ICT. Within this consortium, AI Works is responsible for developing math and reasoning datasets, model validation, and quality assurance. Yoon emphasized, "Our validation capabilities and expertise in data construction can contribute to ensuring the reliability and safety of independent AI models."

As a certified social enterprise, AI Works continues to deliver social value through employment. The company is notable for creating digital jobs for a diverse workforce, including women returning to work, seniors, and people with disabilities. It also operates a separate standard workplace for people with disabilities.

The company is also accelerating its global expansion. It operates a branch in Vietnam, and recently, the Vice Minister of Posts and Telecommunications of Cambodia visited the company to propose collaboration. This includes the development of a large language model (LLM) based on Khmer, the local language, and the training of local talent. Yoon explained, "Khmer is a language that global AI platforms do not support well," and added, "Our aim is to achieve public value by supporting underrepresented languages."

With the explosive growth in demand for AI-based automation, the customer base is expanding to include call centers, the financial sector, and the public sector. Yoon commented, "I believe we are experiencing a second wave after the Digital New Deal," and added, "We are ready." The company has grown to 170 employees and is preparing to raise pre-IPO investment of 15 billion won, with the goal of going public the year after next. He concluded, "We are looking for investors who share our long-term vision and commitment to social value, rather than just short-term financial gains."