Korea’s First Commercial VLM-Powered Video Surveillance Solution
Nota AI
Korea’s Nota AI, a premier provider of AI model optimisation technology, has officially launched its real-time video monitoring solution NVA (Nota Vision Agent). It is powered by advanced generative AI, specifically Vision-Language Models (VLMs).
NVA transcends traditional video monitoring systems that depend exclusively on basic object recognition by analysing intricate scenes, comprehending object interrelations, recognising breaches of operating protocols, and identifying sophisticated risk indicators in real time. Although other firms have presented VLM-powered prototypes, Nota AI is the inaugural firm in Korea—and among the few worldwide—to effectively implement this technology in actual industrial settings.
NVA resolves persistent challenges associated with conventional computer vision (CV) solutions, including recurrent false positives, overlooked detections, and complex rule construction. Its user-friendly interface facilitates natural language enquiries and automated report generation, markedly enhancing operational safety and responsiveness.
After successful field validation, Nota AI has commenced the rapid implementation across many sectors. The company, in partnership with Kolon Benit, successfully executed an eight-scenario Proof of Concept at Kolon Industries’ Gimcheon 2 Plant. A Kolon Industries spokesperson remarked, “NVA enhances oversight by identifying breaches of standard operating procedures (SOPs) and recognising intricate risk patterns prior to their escalation.”
Conventional systems were limited to identifying discrete entities such as “person” or “vehicle,” inadequately addressing the prevention of accidents influenced by complex aspects such as work sequencing, equipment condition, and environmental context. NVA was developed to address these constraints by comprehending real-time situational context. For instance, it may identify multi-factor situations such as “working alone on a ladder without safety equipment” and activate alerts—integrated with field apparatus—within 2.5 seconds via a prompt-based interface.
Rapid implementation and adaptable rule modifications are significant advantages as well. Conventional AI monitoring solutions generally necessitate over three months to establish new detection rules, whereas NVA accomplishes this in under a week. The complete system installation and activation requires about 2 to 3 weeks.
Furthermore, NVA effortlessly connects with current CCTV systems, facilitating straightforward implementation without the need for hardware replacement. The solution accommodates both small sites and large industrial facilities, featuring robust data protection through end-to-end encryption.
Nota AI intends to extend NVA to international markets, encompassing the Middle East and North America, with this launch. The company is presently assessing its applicability in diverse sectors including industrial safety, targeted monitoring, intelligent transportation systems, retail, and childcare, while actively seeking commercialisation through collaborations with both domestic and foreign partners.
“NVA delivers real impact in industrial settings by reducing risks and improving workplace safety,” said Myungsu Chae, CEO of Nota AI. “It represents a practical and effective application of generative AI, purpose-built to address complex challenges across industries. Our goal is to make NVA the global benchmark for industrial AI monitoring.” He added, “As we prepare for our IPO, NVA is more than just a new product — it’s a strategic milestone that showcases our technological maturity and commercial readiness. We believe it will be a cornerstone in establishing Nota AI’s leadership in the industrial AI domain.”
Picture Source:Nota AI