On July 3, the Artificial Intelligence Integration and Application Development Forum, a featured event of the 2026 Global Digital Economy Conference, was held at the China National Convention Center in Beijing. During the forum, the 2026 AI-Powered Application Scenario Showcase was officially released. The jointly submitted case by Changan Automobile and Futong Technology, "Multi-Agent Collaboration for Voice of Customer (VoC) Insights", was selected as one of the representative cases, receiving national-level recognition in the field of the digital economy.


Zijian Zhang (fourth from right), AI Delivery Lead at the Genesis AI Innovation Center of Futong Technology, accepted the certificate on behalf of the joint project team.
Co-hosted by the Beijing Municipal People's Government, the National Data Administration, Xinhua News Agency, the United Nations Development Programme (UNDP), and other organizations, the Global Digital Economy Conference is one of China's most influential international events in the digital economy. Under the theme "AI Empowering the Real Economy, Digital Innovation for the Future," the Artificial Intelligence Integration and Application Development Forum focused on innovative AI applications and value creation across the real economy. During the forum, Beijing released a series of representative AI application cases that showcase the latest achievements in the deep integration of artificial intelligence and industrial development, reflecting the city's systematic efforts to advance the AI ecosystem and build a comprehensive innovation framework.
Case Study | How AI Turns Voice of Customer into Actionable Business Decisions
As the automotive industry accelerates toward intelligent mobility, connected vehicles, and customer-centric operations, customer feedback has become an increasingly valuable asset for product innovation, service optimization, and business decision-making.
To support the group's strategic initiatives and advance its Direct-to-Customer (D To C) operating model, Changan Automobile partnered with Futong Technology to develop the Voice of Customer (VoC) Digital Intelligence Platform. By integrating large language models (LLMs), knowledge graphs, and multi-agent collaboration, the platform enables closed-loop customer insight management that drives business decisions. It connects customer feedback throughout the entire journey—from purchasing and selling to ownership, maintenance, and service—with business processes spanning R&D, manufacturing, supply chain, sales, and operations. As a result, customer feedback evolves from simple data collection into a business-driving capability through an integrated workflow of collection, understanding, routing, and closed-loop management.
AI Capability Upgrade: A Collaborative Multi-Model Intelligence Framework
Driven by Changan Automobile's business requirements, the platform moves beyond the traditional single-model approach by establishing a collaborative framework that combines NLP rule-based models, domain-specific models, and large language models.
LLMs provide end-to-end capabilities including data augmentation, corpus pre-annotation, new terminology discovery, label mining, opinion extraction, intent recognition, sentiment classification, and summary generation. Domain-specific models are optimized for automotive business scenarios, enabling accurate sentiment recognition and intent classification, while rule-based models ensure high accuracy for specific business requirements. Working together, the three-layer model architecture has achieved an accuracy rate exceeding 85% across automotive application scenarios.
Combined with an automotive knowledge graph jointly developed by Futong Technology and Changan Automobile containing more than 10,000 industry knowledge entries, as well as a customer journey tagging system covering purchasing, selling, ownership, maintenance, and service, the platform supports intelligent analysis across customer intent recognition, sentiment analysis, competitor attribution, issue identification, and trend prediction, transforming customer feedback into actionable business insights and decision support.
Data Foundation: Comprehensive Customer Feedback Collection and Real-Time Processing
The platform establishes a unified data collection framework covering both private and public channels.
Private-domain data sources include the brand app and the 400 customer service hotline, with a processing capacity of up to 50,000 records per hour. Public-domain data covers major social media platforms and leading automotive media channels, processing up to 20,000 records per hour, significantly improving operational responsiveness.
Multi-Agent Collaboration Framework
Building upon accurate customer insight, the platform further introduces a collaborative multi-agent architecture consisting of the Intelligent Data Query Agent, Intelligent Analysis Agent, and Intelligent Reporting Agent.
The Intelligent Data Query Agent is responsible for customer intent recognition, semantic understanding, and multi-source data retrieval. The Intelligent Analysis Agent leverages collaborative reasoning across large and domain-specific models to perform sentiment analysis, tag matching, and trend discovery. The Intelligent Reporting Agent generates structured business insights, recommends closed-loop business actions, and supports real-time alert subscriptions.
Through dynamic collaboration among the three agents, the platform establishes a complete operational loop covering proactive monitoring, real-time business response, and post-event evaluation and management.
Business Impact | From Customer Feedback to Business Decisions
The Changan Automobile Voice of Customer Digital Intelligence Platform has been deployed across multiple brands and business units within the Changan Automobile Group, establishing a digital operating framework covering the entire customer experience management process.
In production environments, the platform has achieved 100% intelligent work-order response for core public-domain customer complaints while maximizing the identification and utilization of customer suggestions. It also delivers insight reports within minutes, significantly improving both customer feedback processing efficiency and business decision-making responsiveness.
At the organizational level, the platform has elevated Voice of Customer from a system capability to a management mechanism by establishing a routine customer listening framework for corporate executives and business leaders, enabling continuous awareness of and engagement with customer feedback. Senior corporate management now reviews and evaluates more than 100 customer feedback records each month, strengthening direct visibility into frontline customer issues and management involvement. Business executives also conduct routine reviews of more than 100 customer feedback records each quarter, ensuring that customer voices reach decision-makers directly and creating a governance mechanism centered on listening, understanding, and action.
AI-Powered Customer Experience Management Unlocks Greater Data Value
The selection of this project as one of the representative cases in the 2026 AI-Powered Application Scenario Showcase not only recognizes the achievements of the Changan Automobile Voice of Customer Digital Intelligence Platform, but also further validates the value of large language models, multi-agent collaboration, and industry knowledge systems in enterprise customer experience management.
Looking ahead, Futong Technology will continue to deepen the integration of AI models, industry knowledge, and business scenarios. Leveraging its core products—including the Zhihe Agent Platform and Data Workshop—the company will further embed artificial intelligence into enterprise business processes, enabling scalable and replicable intelligent transformation across industries while continuously unlocking both data value and business value.