Operating telecommunications networks is becoming increasingly complex. With the convergence of fixed and mobile networks, a shift to software-defined infrastructure, and rising demand for customized services, Communications Service Providers (CSPs) face the dual challenge of managing complexity and driving innovation. Huawei Intelligent Operations aims to solve these challenges with a new generation of AI-powered solutions.
The New Telecom Landscape: Challenges in Network Operations
1. Converging Network Requirements
- Modern use cases—like IoT and end-to-end network slicing—require coordinated fixed and mobile infrastructure.
- Reliable service now depends on highly integrated, dynamic environments.
2. Skill Shifts and Operational Evolution
- As networks become software-centric, operators must acquire skills in assurance, maintenance, and service-centric operations.
- Traditional uptime metrics are being replaced by business-focused KPIs and service differentiation.
3. Technology Complexity
- Adoption of enterprise APIs and agentic AI adds new dimensions to managing service experiences.
- New architectures can enable better support for diverse use cases if implemented effectively.
Huawei Intelligent Operations: Strategic Approach
1. Standards and Industry Collaboration
- Collaborates with TM Forum to create service-centric metrics.
- Aligns with industry best practices while pushing AI innovation boundaries.
2. Research & Feedback Loop
- Invests in digital twins and telecom-specific foundation models.
- Feeds operational insights from live deployments back into product improvements.
Real-World Impact: Case Studies from Huawei
Asia-Pacific:
- Used the “Expected Demand Not Served” algorithm to detect service-impacting faults.
- Result: 15% reduction in traffic loss through prioritized fixes.
Middle East:
- Improved topology accuracy in FTTx networks using delay-tolerant architecture.
- Result: 60% reduction in invalid work orders and 10% drop in customer complaints.
Europe:
- Applied Gen-AI with role-based copilots and agent collaboration.
- Result: 25% reduction in Mean Time to Repair (MTTR).
Embracing Gen-AI: Copilots and Agents in Telecom
1. Two Core Models of Gen-AI in Operations
- Copilots: Assist humans with coding, configuration, ticket triage, etc.
- Agents: Execute predefined tasks autonomously, under supervision.
2. Key Challenges to Address
- Hallucinations: Must be mitigated through RAG, verification, and guardrails.
- Cost Efficiency: Requires AI-specific FinOps practices to balance value vs. expense.
- System Integration: Digital twins and Gen-AI interfaces demand robust BSS/OSS data connectivity.
- Data Transformation: Success depends on cleaning, normalizing, and integrating siloed data.
- Training: LLMs need fine-tuning on domain-specific documentation and processes.
DeepSeek: Huawei’s Answer to Gen-AI Optimization
1. Efficient and Effective AI
- Launched in January 2025, DeepSeek proves Gen-AI capabilities can be achieved with greater efficiency.
- Supports domain knowledge construction, code generation, data analysis, and ticket processing.
2. Direct Application to CSP Operations
- Huawei is leading the charge in applying DeepSeek to telecom infrastructure.
- Optimizes performance while reducing the computational burden and cost.
Huawei Intelligent Operations is redefining how telecom networks are managed in an AI-driven world. By aligning cutting-edge technologies like Gen-AI, digital twins, and DeepSeek with real-world operational expertise, Huawei is helping CSPs not only solve today’s complexities but also unlock new revenue streams and service models. With its strategic partnerships, commitment to standards, and hands-on deployment experience, Huawei stands at the forefront of the next generation of telecom innovation.
