With AI now enabling closed-loop self-optimization of networks, the Talent for Tomorrow - Phase II Catalyst is identifying the skills engineers need to redeploy their talents so CSPs can make the most of these increasingly automated networks
Helping humans and robots work together on autonomous networks
As advances in AI enable development of closed-loop network management systems that optimize themselves without the need for much human intervention, telecoms networks are becoming increasingly automated, but also more complex – which will inevitably make some traditional engineering role specifications obsolete. Now then is the time for CSPs to train humans and robots to work together, meaning staff will have to acquire new skills as they turn their focus from performing the maintenance tasks being automated to monitoring the implementation of that automation.
Intent-based, task-based, and knowledge-based closed loops are the key components needed to bring autonomous network to the next level – and if they are to harness the full potential of AI, CSPs need engineers who know how to apply and monitor these new technologies to best effect. But, as a trial-and-error approach to AI product and service development can be expensive and inefficient, CSPs need to develop methodologies to support the deployment of AI in network management.
The Talent for Tomorrow – Phase II Catalyst is exploring human-robot collaboration and robot-robot collaboration in CSPs’ production environments, with participating organizations including Huawei, China Telecom, China Unicom, Broadtech and the China Information Technology Designing Consulting Institute. Based on the methodologies developed by Phase I, the Catalyst is developing a ‘talent transformation architecture’ to support deployment of AI within networks.
In particular, the project is identifying the skills that operational and maintenance engineers need to work with knowledge-based, task-based, and intent-based closed loops across the network, service and business layers. The project referenced TMF IG1252, GB1000, GB1004, GB1007, GB921 in operation, management, and governance to drive autonomous network talent transformation. The Catalyst has developed four new ‘role families’, which describe how to transform cloud network engineers into intelligent network engineers and network innovation engineers, who can create, train, manage, and work with expert robots.
China Unicom has defined three levels for its Talent Maturity Model, which broadly equate to robots assisting humans, robots replacing human functions, and robots surpassing human capabilities:
Level 1: Functional AI model (specific algorithm)
Level 2: Service expert-based LLM (enhanced machine learning + functional expert system)
Level 3: Business expert-based LLM (AI-self learning + scenario-based expert systems)
China Unicom’s expert robot family has currently reached Level 2 of the model and is supporting automated services in 31 provinces.
Applications and wider value
Using this approach, China Unicom has recruited 231 R&D talents and upskilled more than 2,000, thereby avoiding the need for layoffs. The Catalyst also found that effective collaboration between humans and robots can improve operational efficiency by 50%, and quality by 10%, in the service and resource layers of a telecoms network.
Following the deployment of autonomous networks with service expert robots in 31 provinces, China Unicom says the talent transformation architecture contributed to an increase of thousands of million yuan in revenue and a reduction of hundreds of million yuan in network service operation costs. “Currently China Unicom has put eight robots into use,” explains a company spokesperson. “They have saved hundreds of millions of kilowatts of power, and hundreds of persons-year workloads. In general, they have cut costs by billions of RMB.”
The effective deployment of autonomous networks is opening up new opportunities for CSPs and creating more value for their customers. In the enterprise segment, operators can use intent-based, agile network orchestration and support to provide businesses with customized high-quality service experiences, and support broader digital transformation programs. In the consumer market, telecoms operators can use AI to facilitate the development of a new business model where networks develop precise, real-time self-awareness functions, as well as service-based and user-level control features to differentiate experiences.
Autonomous networks could have a particularly major impact in the residential market, as people increasingly need advanced connectivity in their homes to meet both recreational and professional needs, with corresponding challenges to CSPs in terms of reliable service provision. On-demand, fine-grained service experience monitoring and support through AI could be used to guarantee such ‘five-star’ broadband experiences – and with the achievements of this project, we can have confidence that CSPs can meet those challenges by drawing on the strengths of human engineers and AI robots working in harmony, rather than having to choose between one or the other.