Deeply Inc., developer of the industrial acoustic AI solution Listen AI, has announced that its technology achieved 99.87% accuracy in connector engagement sound inspection, addressing one of the automotive industry's longstanding quality control challenges. The company plans to leverage this milestone to expand its presence in North America by participating in Automate 2026, North America's largest robotics and automation exhibition, taking place in Chicago from June 22 to 25.
As the automotive industry accelerates the shift toward Software-Defined Vehicles (SDVs), electrification, advanced driver assistance systems (ADAS), and connected vehicle technologies, ensuring the correct engagement of electronic connectors has become increasingly critical to manufacturing quality and vehicle safety.
As vehicles become increasingly dependent on electronic systems, the number of connectors used across infotainment systems, ADAS technologies, and high-voltage battery architectures continues to rise. Verifying whether these connectors are fully engaged has become a crucial quality assurance requirement throughout vehicle assembly.
Historically, automotive manufacturers have relied on workers to manually identify the subtle engagement "click" produced when connectors are properly inserted. This process depends heavily on human hearing and tactile feedback, creating potential inconsistencies caused by worker fatigue and the absence of traceable inspection data.
Deeply's Listen AI Industrial platform aims to automate this process using acoustic AI technology capable of detecting engagement sounds with greater consistency and precision than traditional manual inspection methods.
According to Deeply, Listen AI has progressed beyond pilot testing and is currently operating in mass-production environments. The company reports successful deployment within the manufacturing facilities of a global automaker, identified as Company H, across production plants in Korea and Mexico.
In these production environments, the solution achieved a connector engagement inspection accuracy rate of 99.87%, meeting stringent quality control standards required in large-scale automotive manufacturing.
Beyond connector inspection, Deeply has completed proof-of-concept projects across battery assembly operations, electric motor defect detection processes, and robotic automated assembly lines. Several of these initiatives are now advancing toward production deployment.
One of the key challenges in automotive manufacturing is the presence of high-noise environments generated by impact tools, air guns, metal friction, conveyors, and other industrial equipment. These conditions often exceed 85 decibels, making it difficult to isolate subtle engagement sounds using conventional audio equipment.
Deeply has developed a manufacturing-focused foundation model trained on more than 10 million process event data points and over 2.1 million hours of factory noise data. This enables the platform's noise-canceling algorithms to filter complex background noise while identifying subtle engagement sounds that may be difficult for human inspectors to detect.
The company also leverages proprietary frequency-band analysis technology to distinguish between Primary Lock engagement frequencies and Secondary Lock frequencies, enabling more accurate verification of connector installation quality.
Unlike conventional machine learning systems that often require weeks or months of on-site data collection and model training, Listen AI utilizes a pre-trained manufacturing foundation model capable of operating immediately after installation.
The solution can be deployed without modifying existing production equipment, manufacturing tools, or assembly line layouts. It supports more than 20 industrial sensor configurations, including directional microphones, wearable microphones, and array microphone systems for high-noise environments.
Inspection results are integrated directly into existing Manufacturing Execution Systems (MES) and programmable logic controller (PLC) environments, enabling real-time monitoring and traceable quality records.
"Connector engagement sound detection was previously considered an area where full automation was impossible, but it is being demonstrated in global production lines that acoustic AI solutions can solve this," said Deeply CEO Suji Lee. "As we are receiving collaboration inquiries from multiple global automakers and parts manufacturers, including those in North America, we will use Automate 2026 as an opportunity to fully enter the North American market, which has complex manufacturing environments, and lead the AI transformation of manufacturing sites."
Deeply will exhibit at Automate 2026 in Chicago, where it plans to showcase live demonstrations of Listen AI's connector engagement sound inspection capabilities.
The company views North America as a strategic growth market due to increasing demand for factory automation, workforce efficiency, and AI-powered manufacturing technologies. Through participation in the event, Deeply aims to establish new partnerships with automotive OEMs and Tier-1 suppliers seeking to modernize quality inspection and production processes.
With proven deployments in automotive manufacturing and expanding applications across industrial quality assurance, Deeply is positioning acoustic AI as a practical solution for improving inspection accuracy, reducing manufacturing defects, and supporting the broader digital transformation of industrial operations.
Acoustic AI solution company 'Deeply' showcases "Machine Hearing" technology that manages safety in industrial sites and daily life through its independently developed non-verbal AI sound analysis technology. By using AI technology to detect sounds that are difficult to identify with human eyes or ears, its application is gradually expanding in public and industrial safety management fields.
'Listen AI Industrial', an industrial AI sound analysis solution developed by Deeply, is a solution that automates product quality inspections that have previously relied on human hearing. By distinguishing even minute differences in component sounds, it is utilized for noise inspection of industrial components such as actuators, motors, and gears, as well as component engagement sound inspection. It is currently being supplied to affiliates of the H Group (an automaker), Hyosung Electric, and Korail. As the physical AI field centered on actuators, such as robotics and e-mobility, becomes more active in the future, the utilization of this solution is expected to increase further.
'Listen AI Safety', a public safety AI sound analysis solution, is an AI solution that identifies safety through sound in restrooms, fitting rooms, and blind spots where human vision or CCTV cannot directly intervene. It identifies unseen danger signals such as screams, loud voices, crying, and requests for rescue, enabling immediate response to safety accidents. It is currently in use at the Government Complex Sejong gymnasium, Naejangsan National Park, Incheon Transit Corporation, and Kangwon Land. It is also being exported to global markets such as the Singapore Ministry of Home Affairs, Thailand, and Vietnam.
TechIntelPro empowers B2B technology decision-makers with the latest industry trends, deep insights, leadership perspectives and actionable intelligence to drive both organizational and professional milestones—and thrive on the tech pulse.