# From surveillance to prevention: Intellivix's AI vision > Source: > Published: 2026-06-30 02:57:56+00:00 Profitable for 13 straight years, Korean safety AI firm is now taking its technology to Japan, Southeast Asia and beyond Choi Eun-soo has spent more than two decades in an industry built on looking backward, but the Intellivix CEO is now betting that cameras can help prevent accidents before they happen. “AI should not be about showing off technology,” Choi said during a recent interview with The Korea Herald at Intellivix’s headquarters in Seoul. “It has to work in the field and prove itself through revenue.” For years, closed-circuit television has served as a record of what went wrong — footage to review after a fire, traffic accident, workplace injury or public safety incident. Choi wants to change that role. His company, Intellivix, is working to turn CCTV from a post-incident monitoring tool into a safety AI infrastructure that detects danger, explains what is unfolding and helps people respond before damage escalates. Founded in 2000, Intellivix has built its business around AI video analytics and safety control systems used in public safety, industrial sites, traffic management and disaster response. Rather than competing in the crowded field of chatbots and general-purpose AI models, the company is focusing on places where missed signals can carry real consequences — factories, roads, railway stations, public facilities, correctional facilities and disaster sites. From monitoring to prevention Traditional video control still depends heavily on human operators watching multiple screens. Intelligent CCTV systems have improved detection, but many still stop at sending alerts. Choi says that is not enough. “When an alarm simply rings, people may be startled without knowing what is actually happening,” he said. “What matters is whether AI can explain the situation quickly enough for people to make the right decision.” Intellivix’s flagship platform, Gen AMS, is a generative AI-based integrated control system that analyzes video feeds, detects risks and summarizes situations in natural language. Rather than simply sounding an alarm, it tells operators where an incident occurred, what kind of risk is developing and what response may be needed. If a fire breaks out, the system can describe the location and severity of the situation and provide guidance based on standard operating procedures. At an industrial site, it can detect a worker collapse, a forklift collision risk or unsafe access to a restricted area, then brief control room operators in plain language. The system combines vision AI with vision-language models, allowing it to read context rather than merely detect objects. In safety control, that distinction matters. AI has to tell an actual fire from a fire video playing on a smartphone screen, or sparks in a steel mill from a dangerous blaze. That push from perception to explanation is also shaping the company’s next products. Intellivix is developing VIXA, a safety AI agent, and Argos, a four-legged patrol robot designed to cover blind spots that fixed cameras cannot reach. A data moat built outside the lab A recurring point in Choi’s argument is that AI that works in a laboratory can fail in the real world. The company says it has accumulated about 750 million domain-specific images from public, industrial, traffic and disaster-related environments over more than two decades of deployment. “A model that performs well in a lab may fail in an actual industrial site,” Choi said. “Steel mills and construction sites have steam, vibration, strong lights, darkness, dust and sparks. Without data from those environments, safety AI cannot work properly.” That field data is central to Intellivix’s case. Crowded transit spaces, construction sites, roads in heavy rain and steel mills glowing with furnace light create visual conditions that general-purpose AI may struggle to interpret. Mission-critical safety applications, Choi argues, require domain-specific AI trained on actual operating environments. Intellivix’s long deployment history, he says, gives the company a data advantage that cannot be built quickly. Taking Korean safety AI abroad Intellivix’s overseas strategy begins with a practical premise: safety AI tested in Korea’s dense and demanding environments can be adapted for other markets. Japan is one of its key references. According to the company, Intellivix has supplied an AI mobility assistance system to Japan’s JR rail network since 2021. The system detects wheelchair users, people carrying white canes and passengers with strollers, helping improve accessibility and safety in stations. “Japan is a very demanding market when it comes to adopting foreign technology,” Choi said. “The fact that Korean AI has been applied to Japan’s core transportation infrastructure is a sign that our technology has been trusted.” Intellivix is pursuing partnerships and distribution channels in Southeast Asian markets — including the Philippines, Vietnam, Indonesia, Malaysia and Bangladesh — with opportunities in urban safety, traffic control and correctional facility monitoring. Australia is another target, particularly for physical AI. Choi said the company is discussing potential uses of Argos there, including airport safety, athletes’ village security and large industrial sites ahead of the 2032 Brisbane Olympics. Beyond robots, Intellivix is also looking to deploy VIXallcam, its all-weather AI vision sensor, in outdoor sites where low light, fog or heavy rain can limit ordinary cameras. Choi said Intellivix aims to generate about 5 billion won ($3.25 million) in overseas sales this year, as part of its broader goal to expand the overseas share to roughly one-third of total revenue. As Intellivix expands overseas, AI video control must also address concerns over surveillance. Choi says the company’s systems are designed to detect dangerous situations, not identify individuals. Features such as facial blurring and restricted video access are designed to keep the focus on safety, with data managed under each client’s own rules. “What we need to know is not who the person is,” Choi said. “What matters is whether that person is in danger.” The goal, he says, is not simply to export software, but to export a Korean model of safety AI infrastructure tested in real operating environments. Profitability before promises Intellivix also stands out for something still unusual in the AI sector: profitability. While many AI companies are built around heavy investment and operating losses, Intellivix says it has recorded profits for 13 consecutive years. According to the company, it posted 46.6 billion won in revenue and 5.4 billion won in net profit in 2025. Its revenue target for 2026 is 70 billion won. “AI has to be proven by customers who are willing to pay,” Choi said. That philosophy also shapes the company’s listing strategy. Choi said Intellivix is preparing for a conventional Kosdaq listing rather than a technology-special listing, a route often used by Korean deep-tech companies with strong technology but limited earnings. “A technology-special listing presents a blueprint of what a company may become in the future,” Choi said. “We are already generating revenue in the market and making money with AI. We want to be evaluated as we are.” Funds raised through the planned listing would be used to advance AI model development and recruit global talent, he said. For Choi, that is where Korea’s AI opportunity lies: not only in building larger models, but in proving that specialized AI can solve real-world problems. “Korean AI can lead through domain-specific deep tech that solves real human problems,” he said. The Top 100 Global Innovators series spotlights the trailblazers shaping Korea’s future across a range of industries — from bold entrepreneurs and tech pioneers to research leaders — whose innovations are making a global impact beyond Korea. More than a celebration of success, the series offers a deeper exploration of the ideas, breakthroughs and strategies driving their achievements. — Ed. yeeun@heraldcorp.com