China Mobile Jiangsu’s AI + BOSS Proactive Service Elevates Customer Experience

Details: Staff Writer; Category: Customer Management; 26 November 2025; 73793 views

Jiangsu, one of the most economically dynamic coastal provinces in eastern China, has a permanent population of over 85 million and has ranked first nationwide in per-capita GDP for 16 consecutive years. Its 13 cities boast diverse cultural heritages and historical backgrounds, creating a “mega-market” with varied demands.

As the largest telecom operator in Jiangsu province, China Mobile Jiangsu has over 60 million mobile users, more than 20 million family users and over 100 million IoT connections. With the market approaching full saturation and fading demographic dividend, the telecom industry’s competition has shifted from relying on traditional resource input and new-user acquisition models toward value-driven, experience-centered operations. Customer experience has become the key battleground for long-term growth.

However, as 5G reaches scale and IT architecture becomes fully cloud-native and intelligent, China Mobile Jiangsu faces four pressing challenges:

1. Misalignment between Services and Customer Experience

Despite offering a wide range of bundles and services, it lacked an accurate mechanism for matching user needs and predicting experience evaluation, leading to confusion during user’s selection. Meanwhile, inaccurate recommendations increased churn risk and led to complaints and unnecessary resource consumption.

2. Limited Service Differentiation

A lack of exclusive value-added services and deep-engagement mechanisms for high-value customers made it difficult to strengthen core user loyalty and prevent potential churn.

3. Fragmented Inquiry & Complaint Handling

Customer-related data was scattered across BSS, OSS and other siloed systems. Complex cross-domain issue identification and troubleshooting rely on manual processes, leading to lengthy handling procedures that significantly affect customer satisfaction and service experience.

4. Rising Complexity in Network & IT Operations

With full-scale cloudification, more than 100,000 annual network change operations required rapid validation that manual processes couldn’t meet. Frequent adjustments increased the risks of triggering signaling storms in the core network. Cross-domain and cross-layer faults (core network, network cloud, and IP bearer network) are on the rise. However, the current O&M system lacks real-time awareness of global topology, performance, and traffic changes, and it takes an average of more than 2 hours to locate a single cross-domain fault. IT O&M still primarily rely on passive response and post-incident handling, leading to slow fault localization and service recovery, which directly affects customer trust and brand reputation.

In collaboration with Huawei, China Mobile Jiangsu integrated LLMs, machine learning, and digital-twin technologies into its AI+BOSS system, establishing a new model of proactive customer service to address these challenges.

1. AI Service Workbench: Delivers Deep User Insights and Empowers Experience-Based Marketing to Achieve Highly Accurate Demand Matching

By leveraging large language model capabilities and integrating five dimensions (basic information, device attributes, App preferences, household services, and consumption characteristics), four evaluations (saturation, slot availability, service bundling, and support needs), as well as family group data, the system enables comprehensive customer insights (over 1 million calls per month), proactive care (over 4 million calls per month), and scenario-based precision recommendations. This has driven a 38% increase in AI-powered recommendation marketing conversion.

2. Intelligent Service Routing for Efficient Closure of Inquiries & Complaints

The intelligent routing engine combines semantic classification with causal-inference models, continuously self-correcting classification accuracy based on historical patterns. Continuous self-correcting improves classification accuracy over time. For simple online enquiries, AI customer service agents provide 24/7 real-time responses. For complex IT issues, an automated ticketing process is newly added to trigger RPA (Robotic Process Automation) completing in-depth queries or repairing actions within 30 minutes, achieving true end-to-end closure.

3. VIP Service Assistant: Integrated “Telecom + Lifestyle” Experiences for High-value Users

For high-value users, the VIP service assistant analyzes consumption patterns and behavioral trajectories, proactively offering communication support and local lifestyle services. In specific life scenarios such as travel or business trips, it proactively offers communication support and local lifestyle services, and provides a one-stop service for booking flights/hotels using exclusive points. This approach achieves an upgraded experience where "service precedes demand," significantly enhancing the precision of care and the perceived value for users.

4. End-to-End BSS/OSS Intelligent O&M for Service Stability

(1) Intelligent Signaling-Storm Prevention

Based on a six-dimensional network simulation model and statistical vector algorithm, it achieves the capability to simulate hundreds of millions of users and perform full-network storm simulations with 90% accuracy. The system can detect impacts, locate sources of impact, and execute emergency response scripts for over 30 scenarios within minutes, forming a closed-loop process of "evaluation-warning-action."

2) Digital Twin Network Engine

Builds a real-time computing framework for data with hundreds of millions of dimensions, automatically collects data to generate dynamic topologies, and integrates multi-source data to construct digital twins. It enables cross-domain alarm correlation and fault propagation modeling, providing real-time visualization of core scenarios such as network status, fault heat maps, and service paths through a unified view.

(3) Intelligent 5GC O&M

Based on the digital twin of network engine, it integrates graph computing, propagation models, and AI algorithms to automatically mine fault relationships and identify root causes, thereby reducing the scope of manual analysis by 80% and shortening the MTTR (Mean Time to Repair) for cross-network domain faults.

(4) Intelligent IT Collaborative O&M

Building a fault perception agent that accurately identifies anomalies using health curves, LLM, and feature-weighted KNN (K-Nearest Neighbour) algorithms; a diagnostic agent that achieves minute-level fault localization through fault maps; and a recovery agent that intelligently repairs 97 types of faults. These three agents work together to significantly reduce the MTTR (Mean Time to Repair) for cross-IT domain faults.

Significant results since commercial launch in January 2025:

Customer Experience Improvements
Package-matching accuracy increased by30%
User decision efficiency increased by40%
VIP user customer satisfaction score increased by 3%

Service Efficiency & Reliability Gains
Overall response efficiency increased by60%
Complex issues closed within 30 minutes
“1-minute detection, 5-minute localization, 10-minute recovery” success rate increasedto 82%

Overall Business Value Enhancement
Net promoter score increased by13%
Complaint rate dropped by 21.5%
Frontline O&M workload dropped by18.7%

Looking ahead, leveraging China Mobile’s AaaS (Ability as a Service) digital capability platform, the project team plans to build a unified, shareable capability framework and replicate China Mobile Jiangsu’s success across the entire China Mobile Group, moving from regional breakthroughs to nationwide empowerment.

The project team will also continue expanding advanced AI and digital-twin capabilities, collaborating with international standards bodies to contribute to global telecom standards. By exporting reusable intelligent-operations methodologies to additional industries, China Mobile Jiangsu aims to help drive the global telecom sector toward a fully intelligent, digital future.