Recently, Zhongshan DPPUL Electronics Co., Ltd. has successively held four AI-themed training sessions in the conference room on the third floor. From September 5th to 12th, 2026, the company will focus on four levels: "AI general Knowledge, enterprise case analysis, personal ability upgrade, and DPPUL pilot Implementation". It will adopt the form of "external experts + project team sharing + internal participation", and be open to regular participants from all departments and employees who sign up voluntarily, to help everyone understand and use AI. To dare to put forward demands and participate in pilot projects. The training adheres to the principle that it can be understood from scratch. It does not require everyone to become an AI technology expert, but rather to learn to entrust repetitive and trivial tasks to AI and leave time for more valuable judgments and growth.
The first session: Break through fear and learn to be friends with AI
On the afternoon of September 5th, the Professor Hui Yong from the project team conducted a special training session titled "In the AI Era, Learn to Be Friends with AI and Don't Let Colleagues Who Are Proficient in AI Overtake You". The course began by addressing common concerns of the employees, helping them distinguish concepts such as large models, prompts, and intelligent agents, and dispelling the concerns that "AI is complex and difficult to understand and non-technical positions cannot learn it". The training focused on introducing the "Prompt Kit + Universal Prompt Card", and demonstrated scenarios such as generating music, product renderings, short videos, and office documents using AI on-site, allowing employees to visually see how AI can assist in writing emails, taking minutes, and organizing requirements. The course also emphasized usage boundaries: which tasks can be delegated to AI, which must be decided by humans, and how to avoid risks such as AI hallucinations and leaks. The training also assigned "7-day implementation mini-tasks", guiding employees to start with one thing this week and incorporate AI into their daily work.
The second session: Case analysis, understanding how AI enters the factory
On the afternoon of September 8th, the project team conducted a training session titled "Analysis of Enterprise AI Application Cases: Understanding How AI Enters the Factory through Real Cases". The main speaker, Cen Junpeng, used real cases from the Jingzhou Factory of Midea, Lenovo Pavilion, Foxconn Nanqing Factory, China FAW OpenMind, and Weichai Group to demonstrate the application of AI in production quality inspection, first-piece confirmation, production scheduling prediction, system question answering, customer service, etc. The training session highlighted four types of tasks that AI excels at: "Visible" - image, appearance, defect recognition; "Questionable" - systems, knowledge, customer question answering; "Calculable" - production scheduling, prediction, scheme optimization; "Stable" - equipment, inventory, risk monitoring. The course also introduced dozens of scenario ideas submitted by 8 departments of DPPUL, all of which expressed the desire to grow together with AI. Everyone gradually reached a consensus: AI does not replace humans; instead, the system provides suggestions, and humans are responsible for confirmation and decision-making, allowing employees to focus on handling abnormalities, quality judgment, and improvement suggestions.
The third session: Upgrade capabilities and build your own AI agent
On the afternoon of September 12th, the third session of the series of training, titled "Personal Capability Upgrade in the AI Era", was held. The Teacher Chen Jiazhe focused on "Enhancing the Lever of Abilities and Expanding the Boundaries of Abilities" and explained the key concepts such as Agent, Skill, MCP, and CLI along with the entry paths through specific tasks. The course used examples like travel route planning and event information analysis to demonstrate how intelligent agents can combine work methods, through calling external services via tool interfaces, to complete information queries, scheme organization, budget verification, and result checks, reducing individual repetitive investment and improving delivery quality. The training pointed out that in the AI era, ordinary people do not need to become technical experts, but can learn to build their own AI intelligent agents: I set the direction, and the intelligent agent implements it for me; first set the standards, and the AI does the repetitive work, while humans make the key judgments and the results are saved for the next time to be more accurate.
The fourth session: Pilot detailed explanation, let AI grow out of DPPUL's real work
On the afternoon of September 12th, the fourth session of "DPPUL Electronics AI Application Popularization Course - Pilot Detailed Version" was held successively. It was presented by the Teacher Huang Qifeng. The course started from the practical cases such as AI quality inspection in Xiaomi smart home factory, the knowledge base of veteran workers from Guangchai Co., Ltd., the initial draft of AI drawing for PC boards from Danfoss, and the AI customer background check by Xiaoman OKKI. It further summarized the four types of tasks that AI usually performs: verification and inspection, writing and organizing, searching and answering, and planning and prediction. Based on DPPUL's actual situation, the training detailedly explained the five pilot directions that the company planned to promote: verification of component packages in the production department, initial draft of PCB in the R&D department, tracking of material requirements in the material control department, new customer information in the business department, and raw material and exchange rate monitoring in the general management office. Each direction was advanced in a collaborative manner following the approach of "setting standards first, AI doing repetitive tasks, humans making key judgments, and results being consolidated and returned", and the five-step pilot method was clearly defined: defining the problem, developing the smallest version, on-site trial use, comparing indicators, and consolidating and reusing. The training also encouraged employees to become "scene discoverers, sample providers, and first-time testers", writing down the most time-consuming and repetitive step of each day, providing de-sensitized drawings, photos, systems, and reports, and using and providing feedback after the pilot was launched.
Let AI assist DPPUL in creating better products, providing better services and winning more customers
The four training sessions build upon each other, progressing from "understanding AI" to "being able to use AI" and finally to "participating in the DPPUL AI pilot project". There are both general knowledge explanations by external experts, as well as real industry cases, and even a practical implementation path tailored to DPPUL's positions. The trainees stated that the training did not cram complex technical knowledge; instead, it explained AI through examples and practical operations, and also made them start to think: Which task at hand is worth trying with AI? Currently, the company's AI platform (Synmogu) is still in the early stages of construction. In the future, it will pilot from high-frequency and easily verifiable scenarios first, gradually forming reusable company capabilities.
AI has already entered manufacturing enterprises. Instead of just observing, it's better to understand and utilize it together. Next, DPPUL Electronics will continue to promote the popularization and implementation of AI applications. We welcome more colleagues to propose scenarios, provide examples, and participate in trials based on real work, so that AI can assist DPPUL in creating better products, providing better services, and winning more customers.
