Customer Service Automation
Guangdong's power-customer service operation needed a more intelligent service backbone for its 24x7 hotline and digital channels so it could improve intent understanding, reduce handling time, and give agents better answers at large scale.
Reported impact
Guangdong Grid improved self-service resolution, shortened handling time, boosted seat productivity, and established a reusable AI customer-service foundation for the energy sector.
Baidu AI Cloud privately deployed a full intelligent-customer-service stack for China Southern Power Grid's Guangdong customer-service center. The project started from scratch with intelligent text service, rebuilt the hotline IVR experience, and added a unified knowledge base so service logic and answers could be reused across channels. At the model layer, Baidu combined intent recognition, entity extraction, multi-turn dialogue, and semantic understanding with Guangdong Grid's own industry customer-service data to fine-tune a power-domain large model. The resulting platform serves both self-service users and human agents: online customers can resolve common requests automatically, while agents receive answer suggestions and structured knowledge support through the same shared service backbone. By combining private deployment, domain data, and self-learning knowledge management, the platform turned a traditional power contact center into an AI-assisted service operation with measurable improvements in resolution and productivity.
Customer requests from online text channels and the 95598 hotline enter the shared intelligent-service layer. The large model classifies intent and entities, the knowledge base returns grounded answers or agent suggestions, and both self-service and human-assisted flows feed conversation history back into the system to strengthen future responses.
Baidu deployed a full set of intelligent-customer-service products and services privately for the Guangdong customer-service center (original: 百度智能云为南方电网广东客服中心提供了私有化部署的一系列智能客服相关产品和服务).
Intent recognition, entity recognition, multi-turn dialogue, and semantic understanding were used to form a power-industry intelligent customer-service large model (original: 通过意图识别、实体识别、多轮对话、语义理解等关键技术…形成电力行业智能客服大模型).
The platform serves 400,000 customers per month, covers 1,200 knowledge items and 36 scenarios, achieves 92% resolution, reduces talk time 10%, and lifts seat efficiency 5x (original: 智能在线客服月均服务客户40万人次…覆盖1200个问答知识，36个业务场景，智能客服整体解决率可达92%…客服坐席接待效率提升5倍).
The project provides intelligent text service for Guangdong Grid's handheld and web portals, upgrades hotline IVR, and builds a unified intelligent knowledge base (original: 提供广东电网掌厅、网厅智能文字客服业务，实现客服中心热线IVR智能化改造，搭建统一智能知识库).
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 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 Baidu AI Cloud Sample May 16, 2025 Guangdong Power Grid Guangdong Power Grid needed to modernize customer-service operations with better intent understanding and a self-improving knowledge base because legacy customer-center knowledge was outdated, hard to search, and unable… Intent Recognition · Knowledge Graph · Multi-turn Dialogue Baidu AI Cloud Sample Apr 28, 2025 State Grid Shandong Electric Power State Grid Shandong needed to make AI practical across customer service, line safety and internal document work in an enormous utility environment where traditional business processes were slow and heavily manual. Common electricity-service handling time fell 70%. AI Middle Platform · BML Full-Function AI Development Platform · Facial Recognition 