Deeply Showcases ‘Manufacturing Anomaly Detection AI Solution’ at AW 2026

In manufacturing sites, “sound” is important data that contains the condition of machines and equipment. However, the human ear has limitations when it comes to precisely and objectively analyzing subtle differences in sound.

Deeply, an acoustic AI solution company, showcased its “manufacturing anomaly detection automation AI solution” at AW 2026, emphasizing “listening automation” for factories.

At the exhibition, the company prepared demo presentations for two representative solutions: “actuator abnormal sound diagnosis” and “connector fastening detection.”

The actuator abnormal sound diagnosis solution inspects abnormal sounds generated by rotating components. It can classify 10 types of defects on mass-production lines to meet market demand from customers who want to identify not only the anomaly itself but also the root causes, such as bearing or gear defects.

According to the company, the solution can replace the traditional final quality inspection method in which workers listen to the sound of finished motors and determine whether they are normal.

The connector fastening detection solution installs a display on the production line and visually notifies workers when fastening sounds are detected.

This solution is in particularly high demand in the automotive industry. As software-defined vehicles (SDVs) become more important and electronic components increase, connectors are used more widely. If connectors are not properly fastened, products may pass quality inspection, but they can later become defective when shocks or other customer usage conditions cause them to disconnect.

Park Han, CTO of Deeply, said, “Automation of sound inspection was impossible before deep learning AI technology advanced. Now AI makes quantitative and objective analysis possible, which is why we have commercialized it.”

He added that beyond the solutions shown in the demo, the company also provides customized solutions for various manufacturing situations, such as gas leak detection and roller wear inspection in steel plants.

Meanwhile, AW 2026, which introduces AI adoption cases across manufacturing operations, closes on the 6th at the entire COEX exhibition hall in Seoul.