As the automotive industry undergoes a dramatic transformation towards digitalization and electrification, Hotai Group (Hotai Motor) has long since moved beyond its role as a mere "vehicle dealership." Instead, it has transformed into a comprehensive MaaS (Mobility as a Service) ecosystem builder, encompassing the "mobility" aspect of "food, clothing, housing, and transportation." Faced with massive and complex owner and service data, how does Hotai break down long-standing internal data silos and truly integrate AI technology into its business operations?
This interview with Hsieh Chin-yun, Minister of Digital Applications at Hotai, and Hsu Shih-ta, Director of the Technology Innovation Office responsible for the underlying technical architecture, explores how Hotai, in collaboration with Google Cloud, has achieved a digital transformation turning point by enabling up to 70% of its employees to be AI-enabled through the implementation of Gemini Enterprise and the construction of a dedicated "AI platform," from the dual development perspectives of business applications and IT architecture.
Spanning 24 years of business history, the urgency of "breaking down data silos" becomes apparent.
To drive digital transformation within a large corporation, it is essential to first deeply understand the pain points of each unit. Hsieh Chin-yun has worked at Hotai for nearly 24 years, gaining experience across various departments including human resources, Lexus marketing, related company Carmax, and public relations and legal affairs. This experience has given her a profound understanding of how serious the "data silos" phenomenon was within the group in the past.
Hsieh Chin-yun pointed out that the automotive industry is characterized by "extremely long consumer ownership periods," with a single car often driven for ten or fifteen years, but the frequency of maintenance visits is mostly only once every six months. This has resulted in Hotai's database in the past being mostly centered on the "vehicle's" repair or warranty history, with very little understanding of the "owner's" actual lifestyle and consumption preferences. The sales unit may know what parts have been replaced on a car, but they don't know whether the owner usually enjoys playing golf, camping, or loves independent travel.
To address this pain point, Hotai launched a major data integration project around 2019-2020, promoting the construction of an internal data center (CDP) and establishing the core strategy of "OneID". By using consumers' "mobile phone numbers" as a unified identifier (OneID), Hotai successfully broke down the barriers between different related companies, deeply integrating data from Toyota, Lexus, HINO new car sales, as well as iRent car-sharing, yoxi taxi dispatch, chicTrip travel planning, Hotai Pay, and other diverse services. This not only allows Hotai to accurately depict car owners' lifestyles but also to proactively and precisely provide customized car purchase, insurance, and rental solutions, realizing true data monetization.
Strategic considerations for choosing Google Cloud and Gemini Enterprise
In the process of promoting enterprise AI, the three major public clouds were actually all under evaluation by Hotai, and the final choice of Google Cloud and its Gemini Enterprise as the core brain was based on rigorous considerations.
Hsu Shih-ta, who is in charge of the technical architecture, explained that the primary strategic considerations are "data localization" and the highest level of enterprise-grade cybersecurity. Data is the core fuel driving enterprise innovation and competitiveness, and is highly sensitive, so it must be fully protected. Gemini Enterprise provides a comprehensive privacy and governance mechanism to ensure that all of HOT's sensitive data assets operate only in a dedicated environment and are never used to train external public basic models.
Furthermore, ease of system integration is a major draw. Gemini Enterprise is "out-of-the-box," enabling secure integration with existing enterprise office applications via built-in connectors. Its intuitive no-code platform design allows even employees without technical backgrounds to easily utilize AI tools, significantly lowering the learning curve.
We insist on building our own "AI middle platform" to empower the scaling of cross-departmental "AI agents".
When implementing AI, many companies tend to directly purchase scattered SaaS solutions on the market, but Hotai has chosen a more forward-looking path: building its own "AI middle platform" internally.
Hsu Shih-ta further pointed out that the core significance of the middle platform architecture lies in the "centralized management of technical architecture and model resources." In the past, if each department purchased different systems, it would not only result in huge technical gaps and maintenance costs, but also make future integration difficult. By building its own AI middle platform, the IT team can uniformly and accurately train and deploy models for different business scenarios such as parts demand prediction and maintenance progress optimization, and allow front-end platforms (such as the internal employee portal) to easily access these AI resources.
With the strong support of the middle platform, Hotai has successfully deployed the application of "AI Agent" to diverse business scenarios within the group, bringing amazing quantitative value:
• Planning and Market Analysis:By using AI-powered proxy tools, the product planning department transforms complex product specifications and related regulatory data into professional AI assistants, reducing data processing time by nearly 30% and allowing employees to focus more on high-value business decisions.
• After-sales repair and upgrade:The service center has initially piloted the use of NotebookLM Enterprise to organize voluminous technical manuals, which has significantly reduced the information search time for frontline technicians by half, enabling them to troubleshoot vehicle problems more quickly and accurately, thus revolutionizing the traditional repair experience.
• Operational and Legal Regulations:Faced with a massive set of group contracts and regulations, the business planning unit used Deep Research's agency tools to conduct in-depth industry analysis, significantly reducing the time to produce the initial draft report from 1 to 2 weeks to 1 to 3 days.
• Education and training and material generation:By combining generative AI technologies such as Veo 3 and Nano Banana, training video materials and question-and-answer databases are generated, successfully saving up to 90% of manual work time.
Confronting the AI illusion, implementing a "human-machine collaboration" culture with a adoption rate of up to 70%.
The introduction of technology is only half the battle; the real challenge lies in reshaping the corporate culture. Currently, up to 70% of employees at Hotai have integrated Gemini Enterprise into their daily core business, a remarkable achievement thanks to the efforts of the dedicated AI Development Bureau.
However, the risks of "hallucination" and operational quality control are the most concerning aspects of generative AI. In response, Hsieh Chin-yun emphasized that the essence of AI is to "empower people." In addition to limiting the scope and authority of AI responses through the underlying architecture at the system level, the key is to educate people and establish a clear human control and review mechanism.
Therefore, Hotai continuously emphasizes to its employees during internal training that AI is an excellent tool that can help quickly organize data, inspire creativity, and produce initial drafts, but it is by no means a panacea. Final decision-making, fact-checking, and ensuring the accuracy of publicly released content still rely on the expertise and experience of human employees. Through this "human-machine collaboration" model, Hotai effectively mitigates the operational risks associated with the illusion of AI.
Towards "AI First" in 2027: A New Milestone from "+AI" to "AI+"
Having acquired a wealth of MaaS mobile ecosystem data, Hotai Motor's future digital blueprint is undergoing a profound transformation. Hsieh Chin-yun stated that the company's previous strategy was "+AI," which meant adding AI as an auxiliary tool to existing workflows; however, the future goal has officially shifted to "AI+," meaning that AI will be regarded as a core infrastructure, and all new services and business models will be driven by AI.
Hotai plans to fully implement its "AI First" corporate vision by 2027, and further plans to integrate advanced cloud resources such as Vertex AI to develop more sophisticated expert systems, such as "financial analysis agents." Simultaneously, by continuously encouraging employees to become accustomed to and work with AI, Hotai Motor is steadily redefining the standards of future automotive retail and smart mobility services, demonstrating its strong ambition to transform from a traditional automaker into a technology mobility giant.
