Park Geun-tae, S2W CTO: “Expanding into enterprise AI, targeting 40% of revenue from non-security business” [fn Person]

[Financial News] “The success or failure of companies’ AI transformation depends less on the performance of AI models than on how accurately they understand an organization’s work context and contribute to solving real-world problems.”
Data intelligence company S2W is extending its data analysis technology, developed in the security sector, into enterprise AI and cultivating non-security business as a new growth engine. The company plans to leverage its “ontology” technology, which connects the meaning and relationships of complex data, to increase the share of revenue from non-security business from about 15% in 2024 to 40% within five years.
“The dark web and hacking forums that S2W has handled since its early days are difficult to analyze because they mix slang with unstructured information,” S2W chief technology officer (CTO) Park Geun-tae explained on the 5th. “We have extended into enterprise AI the ontology and knowledge-graph technologies we developed by analyzing the context of scattered information.”
After working at the Electronics and Telecommunications Research Institute (ETRI) and TmaxCore, Park spent 12 years at SK Telecom Co., Ltd. (SKT), leading work in AI and big data. At SKT, he spearheaded development of the AI agent A., among other projects, and joined S2W as CTO in 2022.
Park sees the competitiveness of enterprise AI as depending on how organically a company connects its data with its actual work environment. The ontology-based decision-making operating system (OS) SAIP embodies this technological philosophy. It is designed to connect data and expertise distributed inside and outside a company so that AI can reflect its work rules and decision-making criteria.
Using this technology, S2W applied ontology and retrieval-augmented generation (RAG) to Hyundai Steel’s internal knowledge platform, HIP, linking about 130,000 documents. It achieved 91% search and answer accuracy and an average response time of one second, replacing the six existing search systems. At LOTTE Members, S2W developed the consumer trend analysis AI service Segment Lab in collaboration with LOTTE INNOVATE. The service can combine spending data from about 43 million L.POINT members with news and social networking service (SNS) information to analyze customer behavior and generate reports.
“SAIP automatically reflects changes in source data in the ontology, and its explainable AI (XAI) technology lets users trace the rationale for analysis results and the factors influencing them,” Park explained. “A key strength is that users can directly verify whether the AI’s judgments are sound.”
Park cited SAIP’s ability to introduce AI in stages, starting with the tasks that need it while keeping existing data infrastructure in place, as a differentiator in overseas markets. “Rather than building an ontology of all enterprise data, it is important to define the problems to be solved and apply the technology in stages,” Park said. “Field deployment engineers (FDEs) design and build systems optimized for each client.”
S2W plans to focus its efforts on advancing SAIP and expanding its market rather than immediately launching new solutions for its non-security business. “Building on the expertise we have accumulated in the security sector and our ability to develop customized AI for businesses, we will develop national security and cybersecurity intelligence, alongside enterprise AX, as our two main business pillars,” Park added.
[email protected] Choi Hye-rim Reporter