Witnessed by 200 Million Gamers! QWan’s Customer Service AI Agent Debuts as an Industry First

文章摘要：As a leader in China’s interest-based social networking sector, TT Voice, a flagship product of Guangzhou QWan Network Technology Co., Ltd., has amassed a user base of over 200 million gamers. The massive user scale, coupled with 24/7 service demands, has put mounting pressure on traditional customer service models. Today, through a strategic partnership with Udesk, QWan has rolled out an AI-powered solution centered on a customer service agent, effectively addressing the pain points in gamer support services.

"Where are the points I just recharged? I need to draw skins right now!" "I can’t log into my account suddenly—my teammates are waiting for me to play together!" "Why did you ban my account? I never used cheats!"...

As a leader in China’s interest-based social networking sector, TT Voice, a flagship product of Guangzhou QWan Network Technology Co., Ltd., has amassed a user base of over 200 million gamers. The massive user scale, coupled with 24/7 service demands, has put mounting pressure on traditional customer service models. Today, through a strategic partnership with Udesk, QWan has rolled out an AI-powered solution centered on a customer service agent, effectively addressing the pain points in gamer support services.

Service Dilemma: Catering to the Needs of 100-Million-Level Gamers

With over a decade of experience in the mobile internet industry, QWan has built a diversified ecosystem covering interest-based social networking, esports, and artificial intelligence. Its service scenarios are characterized by high frequency and diversity. Previously, gamers often encountered two major pain points when seeking support:

1. Rigid Responses: Traditional customer service bots delivered stiff, one-size-fits-all replies. When a gamer reported "recharged points not received", the bot would mindlessly respond "Please check your network connection". When a user complained about "wrongful account suspension", it would pop up a generic FAQ list, completely missing the urgency and frustration behind the gamer’s inquiry.
2. Repetitive Inquiries: For recurring issues like "what to do if recharged points are missing" or "how to appeal a wrongful account ban", traditional bots had a low problem-solving rate, shifting a heavy workload onto human customer service agents.

Enhancing user experience thus became the core priority for QWan to optimize its service ecosystem.

Strategic Partnership: Udesk Customer Service AI Agent

A partnership precisely tailored to address these needs was born. Udesk Technology’s technical expertise and industry experience in AI customer service became the key to resolving Qwan’s pain points.

Leveraging QWan’s business characteristics, the two parties co-developed a customized customer service AI agent, building a service system that achieves full-scenario coverage, round-the-clock responsiveness, and end-to-end intelligence.

This collaboration is not merely a fusion of technologies, but a deep alignment of service philosophies centered on gamers first.

Intelligent Core: Dual-Driven Engine for Efficient Services

1. Intelligent Workflow of the Customer Service AI Agent

The customer service AI agent’s ability to deliver accurate responses to TT Voice gamers’ recharge inquiries and feature-related questions stems from a business-tailored intelligent workflow.

When a gamer initiates a conversation, the system first retrieves the user’s last 10 rounds of historical chat records—for example, whether the user has previously consulted about account binding issues. Next, it activates a knowledge base-matching workflow pre-configured with response logic for high-frequency scenarios such as recharge rules and account ban appeals, enabling precise positioning of relevant knowledge points.

After matching, a code module automatically processes the knowledge point data to clarify whether the gamer is asking about "recharged points not received" or "account unban procedures". Finally, a large language model (LLM) node integrates historical conversations, the current query, and retrieval results to generate a natural, accurate, and context-aware response.

For service exceptions, the agent is equipped with a dedicated exception-handling node: if issues such as knowledge point matching failure occur, the system will instantly push anomaly data to QWan’s Feishu workspace via an API interface, ensuring staff can intervene promptly and avoid prolonged gamer waiting times.

2. Exclusive Knowledge Base of the Customer Service AI Agent

If the intelligent workflow is the "skeleton" of the agent, the business-exclusive knowledge base tailored to QWan’s operations is its "brain". Its core function is to accurately distinguish the diverse needs of TT Voice users, ensuring responses are fully scenario-specific.

When a gamer submits an inquiry, the system simultaneously inputs the user’s question, historical records, source channel, and session ID. After rapid parsing by a code node, the inquiry is routed according to priority: requests requiring human transfer are prioritized the highest, followed by issues related to minor consultations, high-risk topics, etc.

Compliance is strictly enforced: Sensitive prompts are automatically added to high-risk topics, while scenarios involving minors, police-related issues, and complaints are mandatorily transferred to human agents. During late-night hours, before transferring to human support, the system proactively informs gamers of the official working hours to prevent unnecessary waiting.

Meanwhile, the AI is precisely constrained: response scope, templates, and formats are all limited to the content within the knowledge base. The agent is strictly prohibited from making unauthorized compensation promises or leaking prompt words. It then retrieves the exclusive knowledge base to match relevant points, and finally determines whether to transfer to a human agent based on the results, ensuring precise and compliant service delivery.

Remarkable Results: Dual Leap in Experience and Efficiency

Since the launch of the co-developed customer service AI agent, QWan has achieved a dual leap in gamer experience and service efficiency—a transformation powerfully validated by data.

For TT Voice, QWan’s core product, the agent has achieved an impressive daily conversation matching rate since its deployment.

From "basic problem-solving" to "pursuit of ultimate user experience", the collaboration between QWan and Udesk has reconstructed the core logic of gamer services through the customer service AI agent. It not only injects new momentum into QWan’s own development but also ensures that every gamer’s needs are addressed in a timely manner. This is a vivid practice of QWan’s mission: "To make sure no gamer feels alone in the world".

The article is original by Udesk, and when reprinted, the source must be indicated：https://my.udeskglobal.com/blog/witnessed-by-200-million-gamers-qwans-customer-service-ai-agent-debuts-as-an-industry-first.html