NexCOBOT Auto-training uses Microsoft Cognitive Service, which enables customers to train their own AI model to recognize objects in 10 minutes without the support from data scientists or machine learning experts. However, taking photos and training all products requires a lot of time and effort which is a challenge for business expansion. It is also very difficult to control the image quality between different operators. NexCOBOT's Auto-training machine can automatically take images of each product, while simultaneously integrating the data into the customer's algorithm. this accelerates the object training process and minimizes the human errors. This demonstration showcases how a robot can take images automatically while auto-labeling software can help generate raw picture data with labels. With this solution, tremendous time and effort can be saved for AI training.