On September 5, the Automotive Home AI Product Launch Conference was held in Beijing. With the theme of "car things, find cheese", this conference focused on AI to reconstruct a new ecosystem of car services, and officially released the "cheese car steward", an AI agent product of the Automobile House. The event brought together AI technology eco-partners, new automotive retailers, used cars, car service merchants, industry media, and vehicle owners' representatives to witness the strategic landing of the Automotive House in the field of AI automotive vertical services, and to explore the changes and opportunities brought by artificial intelligence to the entire life cycle of automotive consumption. Liu Shu, chairman and CEO of the on-site Automobile House Board of Directors of the Automobile House "Cheese Cart Manager" conference, said that the changes brought about by AI are not only the upgrading of technical capabilities, but also the way automobile services are being redefined. Over the past 20 years, AutoHome has accumulated a wealth of experience in automotive content, data and user services, and in the new technological environment, these accumulations are being further translated into the ability to solve users' actual problems. Faced with the practical problems of increasingly complex model selection, increasingly scattered consumer information, and numerous entrances to car services, the launch of Cheese Cart Butler is not only a product-oriented practice for the automobile home to combine long-term industry accumulation with AI capabilities, but also a key strategy for the AI + car service track. Liu Zhu, Chairman and CEO of the Board of Directors of Automobile House, made a speech on AI solution efficiency. The difficulty of real ecology supplementing trust and compliance with automobile consumption is changing. Vehicle models and configurations are not scarce. What is really scarce is credible judgment; there are not many service resources. The real trouble is that users do not know who to look for, how to choose, and whether the results are reliable. CTO Xiang Bibo summarized the logic of the cheese cart steward into two layers: selecting the car to solve the "trustworthy decision" and using the car to solve the "reliable service". Looking further ahead, after the gradual landing of advanced intelligent driving such as L4, the car will also move from "waiting for people to initiate services" to "autonomous identification and call services", which puts forward new requirements for the automotive service infrastructure. In the matter of car selection, AI has the advantage of processing information, but more information does not mean that the answer is more credible. The high convergence of models at the same price level, the mixed marketing content and real experience, and the complex price system make it more difficult for consumers to form a clear judgment. Based on this, AutoHome proposed that "cheese selection agent = AI ability + real owner ecology": AI is responsible for understanding needs, integrating information and improving decision-making efficiency, and real owners use first-hand experience to supplement real scenarios that cannot be covered by public information. For users, the goal is very direct - choose a car and ask for cheese, use the car to find cheese. Another difficulty in releasing AI strategic vehicles is the decentralized service and high reliance on trust. Maintenance, repair, charging, insurance, rescue and other needs are low-frequency and complex, and users often have to repeatedly compare between multiple platforms and merchants. In response to this problem, another set of capabilities of cheese cart stewards is "AI capability + service ecology": AI first understands the needs, and then connects with service resources that can actually perform. Therefore, "using a car to find cheese" is not a generalized slogan, but hopes to compress information judgment, service matching and actual undertaking into the same path as possible. Facing the L4 era, this set of capabilities will further extend from "service car owners" to "service car machines", reserving interfaces for scenarios such as autonomous charging, cleaning, repair and maintenance of vehicles. From car selection judgment to car service, AI began to enter the product design of real car life cheese car stewards, which also continued this idea. Rather than positioning itself as a “more chatty” car Q&A tool, it tries to get into real decision-making and service scenarios. Chen Xi, head of cheese cart butler products, said that the ability of the AI agent does not depend on how many questions can be answered, but whether it can give users a clearer judgment on car selection and continue to undertake services when subsequent needs appear. For example, when users hesitate between two models, cheese will not stop at parameter comparison, but will put together conditions such as space, battery life, refueling, and home use, give tendentious suggestions, and explain the reasons for choice. The logic behind this is to reorganize the model parameters, professional testing and real owner feedback into information that is "useful to this user". This is more important than simply increasing the number of answers for a category as high a decision cost as automobiles. After Chen Xi, the product leader of the cheese cart steward, explained the product capabilities, the problem did not end. When buying a car, the market check, discount matching, program calculation, and maintenance, insurance, maintenance and other needs after buying a car often require users to find the entrance again. Cheese cart stewards let these needs originate directly through natural language, and then the system completes the judgment and resource matching. The product does not deliberately emphasize every trading link, but closes the user's most perceived value into two things: less tangling when choosing a car, and less worry when using a car. Supporting this experience is the dual base of the "industry model × industry ecology". The large model of the automobile industry developed by the Automobile House undertakes demand understanding, vehicle knowledge and auxiliary decision-making; brands, dealers, and industry resources such as finance, insurance, maintenance, and repair are responsible for truly bringing services offline. The former solves the problem of "figuring it out", the latter solves the problem of "catching it". This is also one of the core differences between automotive vertical AI and general AI: the value is not only in the answer itself, but also in whether there are enforceable services after the answer. At present, the cheese cart steward has also accessed and launched the three major platforms of Qianqian, Xiaoyi and WorkBuddy to further expand the service portal of AI automotive services in different intelligent ecosystems. Ecological co-construction: two-wheel drive, build a new ecology of collaborative services If there is only a model, there is no continuous update of real experience and a delivery service network, it is easy for the AI agent to return to the "only answer" state. Cheese cart stewards therefore put the user ecology and the merchant ecology in the same important position: one end is responsible for providing a trusted experience, and the other end is responsible for catching the demand. Liu Lei, head of cheese cart butler operations, said that real car owners are an important source for the continuous growth of AI capabilities and an important basis for cheese cart butlers to build trustworthy service capabilities. In the face of the large amount of information in the automotive industry and the relative scarcity of real car experience, the platform relies on the foundation of 150 million stock car owners, and continues to precipitate the first-hand experience accumulated by car owners in car selection, car purchase and daily use as an important source of AI knowledge, and builds a three-tier driving force of "rights and interests, interaction, and honor" around participating users: through car rights and interests and special incentives to share and get returns, through fun interaction to enhance the connection between car owners, through the honorary level system to make long-term contributions can be seen, and the head contributors have the opportunity to get free cars and other rights in the future. Under this mechanism, users can either learn from the real owner's experience to complete consumption decisions, or share their real experiences after becoming a car owner, forming a continuous cycle of "gaining experience — contributing experience". At present, the ecology has precipitated 1 billion authentic reviews and 50 million car owners' word of mouth, and a large number of real questions and answers and car owners' experience continue to iterate on the ability of cheese car butlers, making the real car owners' ecology further become an important support for AI to establish professional judgment and user trust from the content supply. Liu Lei, the head of cheese cart butler operations, shared that the other end of the user ecology is the merchant ecology. Zhao Hongliang, head of cheese cart butler merchant ecology, said that the car home has connected nearly 30,000 cooperative 4S stores, more than 12,000 used car dealers and more than 100,000 home car service resources, covering 369 cities across the country. For the platform, the scale of resources is not the ultimate goal. What really matters is whether the merchants can be screened out and the services can be stabilized. Only when the user's needs can be reliably accepted can AI truly enter the automotive service. Zhao Hongliang, head of cheese cart butler merchant ecology, explained the merchant ecology upgrading strategy. To this end, the cheese cart butler has established access mechanisms such as qualification verification, ability verification, and credit rating, and has incorporated user evaluation and agent performance verification into service scoring. Whether the price is consistent, whether the service is on time, and whether the promise is fulfilled is the basis for evaluation. High-quality merchants get more exposure and users, while low-quality services are reduced or even withdrawn. At the same time, merchant agents such as Auto House Car Merchant Hui, Car Zhi Ying, Lingji, Yuan Buy, and Yuan Guan assume different business and service roles. The core of it is not to add more tools, but to make high-quality supply more easily recognized and invoked by AI, ultimately making users more worry-free and allowing good merchants to obtain more definite growth. The roundtable forum of the conference focused on how AI really enters the industry and life. Xiang Bibo, CTO of the Automobile House, Vice President Tian Chengbei and AI eco-partners, cross-industry representatives, as well as practitioners in the fields of automotive new retail, used cars, and car-using services. From shopping, office, health to automotive services, AI agents are shifting from "giving answers" to participating more deeply in the service process; in the automotive industry, this change further falls into specific links such as car selection, trading, and car use-which user pain points can be resolved by AI, and which service chains can be reorganized, which has also become the focus of discussion. Combined with the practice of insurance, residential and other industries, the two dialogues talked about the evolution of general technology to the landing of the vertical scene of the car and reflected each other, which also provided a more industrial perspective for the product positioning of the "one-stop" car service of the cheese cart steward. That's what a cheese cart butler means to a car home. It is not simply moving existing car information into AI, but trying to reorganize the relationship between "information — judgment — service": using AI to improve efficiency, using real car owners' experience to enhance credibility, and then using high-quality merchant ecology to complete service acceptance. In the future, this set of capabilities will continue to extend to more automotive life scenarios, but for users, the most important thing to remember is not complicated - choose a car to find cheese, and use the car to find cheese. About the car, find cheese.