Gasgoo Munich- "More capabilities don't necessarily make a factory smarter." That was the message from Tan Yingcong, Changan Automobile's deputy chief engineer, on July 3. Speaking at the "2026 Embodied Intelligence Industry Scenario Integration Conference," hosted by Gasgoo, Tan argued that while AI adoption in manufacturing is surging, it remains largely fragmented—building one model per scenario, one app per problem, one project per plant. The critical next step, he says, is shifting from isolated intelligence to holistic intelligence.
Tan Yingcong, Deputy Chief Engineer of Changan Automobile
Tan outlines four stages of manufacturing evolution: automation tackled efficiency, digitalization handled connectivity, and AI addressed cognition. But the factory of the future requires a fourth phase—holistic intelligence. In this stage, domains spanning quality, equipment, energy, processes, and logistics must synchronize, embedding real-time value directly into business workflows.
The urgency is driven by three trends reshaping auto manufacturing: production is becoming increasingly flexible, with rapid switching between multi-model, small-batch runs becoming the norm; system coupling is intensifying; and decision-making demands ever greater speed.
Yet, building the factory of the future means bridging four critical gaps. First, the data gap: information on equipment, quality, processes, and production is scattered across disconnected systems. Second, the knowledge gap: on-site experience, process rules, and judgment standards remain trapped in the minds of veteran workers or in physical documents. Third, the process gap: AI analysis results struggle to flow naturally into business judgment workflows. And fourth, the execution gap: once an AI makes a call, determining the physical mechanism for execution remains a challenge. "Efficiency issues in AI implementation are widespread today," Tan noted.
Tan breaks down the capabilities of the future factory into three pillars: autonomous operation, where systems sense site conditions in real-time to drive execution; autonomous collaboration, where domains and systems link across boundaries rather than operating in silos; and continuous evolution, where every action refines knowledge and optimizes processes. Combined, these form what he calls "intelligent agents."
Tan stresses that a factory agent is not merely a chatbot, a large language model, or an RPA process. It is an intelligent unit designed for manufacturing tasks. It must understand objectives, organize cross-domain knowledge, sync with real-time data, and integrate with business workflows—ultimately driving digital systems or physical equipment to turn decisions into action.
Changan is exploring this across four domains: quality, equipment, energy, and production. In quality, where defect attribution involves complex factors like process, equipment, materials, personnel, and history, agents can recommend improvements and track implementation. For energy, they provide holistic diagnostics by combining temperature, humidity, lighting, production load, and usage strategies. In production, agents coordinate bottlenecks, deviations, anomalies, and materials to enable real-time identification and dynamic adjustment.
Looking ahead, Tan proposes a "platform-data-agent trinity" roadmap, rolled out in three phases: first, laying the foundation and closing the loop on single scenarios; next, expanding scenarios and optimizing models to scale from points to areas; and finally, establishing a decision hub for multi-directional autonomous decision-making. "Agents will increasingly act like digital employees or experts in the factory," Tan said. "They will need to collaborate with humans but also judge independently—helping factories sense issues faster, make quicker decisions, and execute with greater stability."
