Award Category: China Automotive New Supply Chain Top 100
Field: Smart Cockpit
Technology: AI Agent
Key Innovations:
Anchored by an AI voice "central brain," the system bridges mobility and entertainment to power high-frequency applications like interactive podcasts, AI calls, pet wallpapers, AI mini-games, and parking memory.
The solution’s standout feature is its break from the traditional "remote control" approach. It truly understands human language rather than forcing users to adapt to the machine. It supports one-click orchestration of complex intents—no need for users to break down tasks manually. The cockpit boasts cross-session memory, ensuring a seamless user experience, and supports personalized settings via an edge-cloud architecture that evolves with computing power. Additionally, full-link streaming processing brings AI call experiences remarkably close to human interaction.
For automakers, AI Agent slashes R&D costs, sparing them the headache of simultaneously building algorithm teams, curating training data, securing computing power, and handling automotive-grade adaptation. It eliminates the need to start from scratch. At the same time, it compresses the technology window, allowing OEMs to keep pace with large-model iterations while sidestepping common risks in mass production delivery.
For end users, this means an end to functional silos and disjointed interactions. The central brain coordinates navigation, vehicle control, and media systems globally, enabling multi-task linkage with a single command. Full natural language understanding makes the system instantly usable—no need to memorize standard commands. Long-term user profiling and cross-session memory ensure the experience resumes right where it left off, while the connected Agent offers flexible expansion for information queries.
Technical Description:
- Hybrid Model Architecture. Leveraging a multimodal large model framework, this system builds an end-to-cloud collaborative hybrid inference engine. The bottom layer taps vehicle bus protocols for precise data collection and command execution. The middle layer handles spatiotemporal alignment and feature fusion of multimodal data. The upper layer relies on cloud large models for complex scenario reasoning and long-term memory storage. The on-device model slashes local response latency to under one second while ensuring the system grows smarter with continuous interaction.
- Multi-Dimensional User Profiling. By integrating driver biometrics (such as emotional state and fatigue levels), behavioral patterns, and environmental factors (like location), the system constructs a dynamic cognitive model and personalized user profiles. A time-series analysis engine tracks behavioral evolution, supporting cross-scenario intent prediction. This allows for automatic pushes like route optimization suggestions during commutes or smart travel recommendations.
- Humanized Decision Engine. It interprets contextual scenarios and user intent priorities to generate human-like interaction strategies. This decision engine is tightly integrated with the vehicle control system.
Covering core scenarios such as commuting, leisure, and work, including:
- Intelligent Driving Steward
Our Intelligent Driving Steward monitors driver status in real time, detecting fatigue or abnormal risks. When a risk is identified, it proactively responds by suggesting rest stops, playing music, or adjusting cabin lighting to mitigate danger. This shifts the paradigm from simple risk warnings to active safety support, enhancing both driving safety and comfort.
- Natural Language Vehicle Control
Our AI agent leverages natural semantic analysis and cross-functional collaboration to precisely capture user needs—such as improving cabin air quality by automatically triggering coordinated actions from the air conditioning and purification systems. This represents a leap from single-command execution to proactive, multi-task scenario management, boosting both convenience and safety.
- AI Call Secretary
Our AI Call Secretary interacts via natural voice dialogue to help users complete tasks like restaurant reservations, appointment scheduling, and itinerary management. By intelligently understanding needs and adapting to changes, it delivers a more efficient and convenient daily experience.
Launched by Gasgoo, the "Golden Gear Awards" uphold the mission of "discovering great companies, promoting great technology, and empowering industry professionals." The 8th edition in 2026 adopts "China Solutions for the World" as its annual theme, focusing on core automotive technologies and innovation tracks. It seeks technologies, products, engineering systems, and industrial collaboration capabilities formed in the Chinese market, validated by the industry, and possessing global replication value. Centered on real-world problems, industry validation, global adaptability, and long-term value, the awards will leverage Gasgoo’s supply chain database of 320,000 companies and its research capabilities. Through multi-dimensional evaluation including corporate submissions, industry data, and expert review, the awards aim to uncover innovative achievements with industrial value and global potential, further consolidating them into an innovation case library for the global automotive industry.
