The rapid rise of Generative AI (GenAI) has created a wave of excitement and speculation across the industry. I am often asked, “How can GenAI complement existing investments in cloud, automation, and traditional AI to accelerate the path to autonomous networks?”
For communications service providers (CSPs) in Asia grappling with the complexities of 5G deployment and digital transformation, the issue is not the technology’s potential but its practical application.
The answer lies in a synergistic approach that brings together the strengths of GenAI and traditional AI. There is a golden opportunity for CSPs to combine traditional AI and GenAI capabilities by integrating their current AI/ML-driven insights with large-scale language models (LLMs) that are specifically designed and fine-tuned for the telecom industry.
While data and intelligence are not new concepts for CSPs or the telecom industry as a whole, this movement is a big step forward from the data- and insight-driven era of the past. For over 50 years, service providers have used basic data to create programs that could take specific actions. Over the past 15 years, this data has evolved into telemetry data that is used to model insights into actions.
In this era of “knowledge,” I believe CSPs will begin to move toward the ultimate vision: Level 5 autonomous networks, as defined in the TM Forum’s Autonomous Network Maturity Model. At this level, CSPs can execute intent-driven, autonomous operations to deliver zero-wait, zero-touch, zero-trouble services to meet the changing needs of their customers.
Beyond the hype
CSPs can take their first steps by moving beyond the GenAI hype and integrating traditional AI with GenAI capabilities. This move requires a pragmatic approach, and CSPs should look for partners with deep domain knowledge that can help them take several key steps to close the AI gap. This includes:
Curate a vast collection of federated knowledge sources: Leverage unmatched deep domain knowledge across network planning, design, operations, and customer care to power productivity gains and future product enhancements. Develop advanced LLMOps system to optimize knowledge sources: Leverage cutting-edge LLMs designed to maximize the benefits CSPs can derive from their knowledge sources. Evaluate the use of leading LLMs from the open source community that are powerful, task-relevant, and cost-effective to deliver customized outcomes to CSPs and enterprises. Implement domain-driven hallucination management: Adopt prompt engineering and fine-tuning techniques for telco-specific LLMs to achieve consistency and accuracy in AI-generated output. This will reduce the risk of poor quality output by implementing domain-driven hallucination management. Prioritize upskilling of existing talent on data vectorization tools, fine-tuning techniques, and prompt engineering skills.
As CSPs, many of whom are in the midst of rolling out 5G and other cloud-based transformation projects, have concerns and questions, and they want to know whether they should wait until GenAI evolves and matures.
The answer is no. There is no need to wait. Innovative partners are already delivering measurable benefits today to CSPs willing to make the leap to GenAI. Imagine a large Tier 1 CSP looking to realize significant cost savings and efficiencies in their network and service operations. This CSP submits multiple Requests for Information (RFIs) and selects a partner that can guide the CSP personnel along the traditional AI/GenAI blueprint above. They then achieve real, measurable KPIs such as an 80% reduction in knowledge acquisition time, a 72% increase in data analytics efficiency, and over €6M in annual savings in network operations.
This success will be leveraged to create a future-proof blueprint for large-scale automation with GenAI to further improve high-impact areas such as auto-generation of network design configurations and optimization recommendations. It all starts with integrating today's traditional AI/ML insights into the telco-trained GenAI LLM.
Pioneering examples in Asia
Asia is at the forefront of technological innovation, with several countries already demonstrating the power of combining GenAI with traditional AI in network operations.
For example, one pioneering example in Asia is China Mobile's efforts to implement AI-driven network automation, which has reduced knowledge acquisition time by 80%, improved data analysis efficiency by 72%, and saved millions of euros in network operation costs annually, according to a TM Forum report.
Additionally, Asia-Pacific telecom giants such as Singtel and SK Telecom are making great strides in integrating AI technologies to enhance network automation. Singtel's implementation of AI-driven predictive maintenance has improved equipment uptime by 10-15%, and McKinsey research suggests that predictive maintenance could save global manufacturers $240-630 billion.
Examples like these highlight tangible results that address CSPs’ most pressing business challenges and underscore the potential of integrating GenAI with traditional AI in enabling autonomous networks.
Undeniable potential
While the potential benefits of GenAI are undeniable, it is important to approach this technology with a pragmatic mindset.
CSPs should avoid the temptation to wait until GenAI matures before taking action: by partnering with an experienced provider, they can start realizing tangible results today.
Consider a large Asian CSP looking to improve network efficiency and reduce operational costs. By working with a partner to implement a combination of GenAI and traditional AI strategies, they can realize significant improvements in areas such as:
Accelerated knowledge acquisition: Reduce the time spent searching for relevant information Enhanced data analytics: Gain deeper insights from network data Optimized network operations: Deliver significant cost savings
The early successes we have seen have served as stepping stones to more ambitious efforts, such as automating network design and configuration. Of course, the journey to autonomous networks is complex and multifaceted.
Going forward, Asian CSPs will need to understand the interplay between GenAI and traditional AI and leverage the expertise of trusted partners to position themselves as leaders in this transformational era. I believe the benefits in operational efficiencies, customer satisfaction, and new revenue streams will be immense.
The views expressed in this article are those of the author and do not necessarily represent The Fast Mode. The information provided in this article has been obtained from sources that The Fast Mode believes to be reliable, but The Fast Mode cannot be held liable for any loss or damage resulting from any limitations, changes, inaccuracies, misstatements, omissions, or errors in the information contained therein. Headings are for ease of reference only and do not affect the information presented.
