Dipli Launches 'Listen AI Industrial,' an AI-Powered Acoustic Inspection Solution for Manufacturing.

Quality inspections in manufacturing industrial sites have often relied on the hearing of skilled workers, leading to persistent problems of subjective judgment and lack of consistency. Acoustic AI solution startup Deeply has launched 'Listen AI Industrial,' an AI acoustic inspection solution designed to overcome these limitations and revolutionize quality control. 'Listen AI Industrial' is centered around Deeply's unique AI technology that converts subtle 'sounds' generated by mechanical defects into quantitative data. Without needing to stop equipment operation or make contact, it maximizes the advantage of 'hearing' to detect subtle changes in equipment status through sound. This solution directly performs acoustic inspection (abnormal sound inspection), which was previously the domain of skilled workers, using artificial intelligence. It analyzes sound patterns generated by key components such as motors, bearings, and gears in real-time, accurately determining whether they are normal or defective with a numerical judgment within 0.2 seconds. This innovatively reduces inspection variations caused by human intervention, such as worker fatigue, shift work, and environmental noise, significantly enhancing the consistency and reliability of QC (Quality Control) and QA (Quality Assurance) processes. It has been applied to production lines of major manufacturing companies both domestically and internationally, demonstrating over 99% inspection accuracy and the effect of reducing inspection time. 'Listen AI Industrial' has expanded its analysis range beyond the audible band to include the ultrasound band, capturing early defect signs, and is equipped with a self-learning algorithm. Furthermore, for manufacturing environments where security is critical, it is designed as an on-premise solution instead of cloud-based, allowing for independent operation, and it flexibly integrates with existing NVH (Noise, Vibration, Harshness) or EoL (End-of-Line) inspection workflows. Park Han, CTO of Deeply, emphasized, "The core is to quantify pass/fail inspections that relied on human ears, thereby ensuring quantitative reliability and reproducibility. 'Listen AI Industrial' is a solution that dramatically improves accuracy and speed by automating the non-visual aspects of quality inspection." Starting with 'Listen AI Industrial', Deeply plans to expand its technological domain into a 'Physical AI' platform that digitizes the senses of the factory floor. Lee Suji, CEO of Deeply, stated, "Deeply envisions a future for manufacturing AI where AI becomes the 'ears' of the manufacturing process, and further evolves to encompass the judgment and control of the entire process." She expressed the ambition to open new horizons for manufacturing AI beyond quality inspection.