Customer Service Automation
37 Interactive Entertainment's legacy intelligent customer service stack struggled to understand gaming slang and fast-changing subculture knowledge, forcing agents to search fragmented knowledge bases manually and miss service-level targets.
Reported impact
The rebuilt assistant delivered more human-like answers, reduced manual escalations on unclear player requests, and lifted SLA attainment from 61% to 99.3% while improving complex-issue diversion accuracy.
37 Interactive Entertainment rebuilt its game-support stack on Baidu AI Cloud using large models and retrieval-augmented generation. Qianfan ModelBuilder was used to tune prompts and semantic understanding for gaming slang and subculture terms, while Qianfan AppBuilder orchestrated RAG flows that combined the existing service knowledge base with player behavior data and a dynamic knowledge graph. Baidu AI Cloud Keyue Intelligent Customer Service became the frontline support layer, binding common intents such as password recovery or event-rule questions to standard prompt templates for fast answers. More complex issues, including bug and recharge cases, triggered multi-turn dialogues to collect device and scenario details before the system called technical-team interfaces to create tickets. Unresolved conversations were clustered and reviewed so prompts, constraints, and tone controls could be updated continuously.
Player questions enter Baidu AI Cloud Keyue, where Qianfan AppBuilder routes them through RAG workflows that retrieve the customer-service knowledge base, player behavior data, and the dynamic knowledge graph. Qianfan ModelBuilder-backed prompts then generate answers, while complex issues trigger multi-turn clarification and ticket creation through technical-team interfaces.
37 Interactive Entertainment rebuilt its intelligent customer service system with Baidu AI Cloud large models plus RAG technology (original: 三七互娱基于百度智能云大模型+RAG技术重构智能客服系统).
Complex-issue diversion accuracy increased by 2.4x and SLA attainment rose to 99.3% (original: 实现游戏垂直领域语义理解准确率从62.3%提升至行业领先水平，SLA达成率跃升38.3个百分点至99.3%).
High-frequency questions are bound to prompt templates for second-level responses, while complex issues trigger multi-turn dialogues and technical-team ticket interfaces (original: 将高频问题…与提示词模板绑定，实现秒级响应；对复杂问题…再调用技术团队接口生成工单).
Service SLA attainment improved from 61% to 99.3% (original: 服务SLA达成率从61%提升至99.3%).
The model combines customer-service domain knowledge with player behavior data to build a dynamically updated knowledge graph (original: 将游戏客服领域的专业知识…与玩家行为数据结合，构建动态更新的知识图谱).
The system adds empathy constraints such as 'player-centered' and 'friendly tone' so AI responses balance professionalism and affinity (original: 通过添加“以玩家为中心”“语气友好”等提示词约束，平衡AI回答的专业性与亲和力).
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
Baidu AI Cloud Sample Jan 8, 2026 China Mobile Online (中移在线) China Mobile Online needed faster and deeper customer-intent insight than manual sampling and call listening could provide so it could support precision marketing, real-time service intervention, and low-threshold model… Clustering analysis timeliness is reduced to less than 1 hour from about 3 hours. Customer-Agent-Business Profiling Engine · Real-time Intelligent Insight Platform · Real-time Knowledge Push Engine Baidu AI Cloud Sample Jan 8, 2026 Mercedes-Benz Group AG Mercedes-Benz needed to unify fragmented customer-contact operations spread across assistant, telematics, and roadside-assistance centers so it could improve oversight, expand service capacity, and deliver a more… End-to-end online Q&A accuracy exceeded 85%. ASR · Baidu AI Cloud Intelligent Customer Service · NLG Baidu AI Cloud Sample Jan 8, 2026 State Grid Corporation of China State Grid's 95598 service center faced heavy call volumes, limited channels, and rising pressure during demand peaks and outages, making it difficult to maintain service quality and scale customer support efficiently… Average daily interaction volume is 25 million messages. AI Base Platform · AI Operations Platform · Baidu AI Cloud Intelligent Customer Service 