A co-existence story: GenAI expands Traditional AI’s horizon

Introducing Generative AI (GenAI) to your Telco AI strategy will be a game-changer. It not only speeds up knowledge discovery and content generation but can also be applied to a wide range of tasks such as alarm intelligence, ticket resolution, network security, network planning, and service management automation. GenAI enhances autonomous decision-making with assistive reasoning, bringing Level 4 and Level 5 autonomous operations closer to commercial reality for communications service providers (CSPs).

Reaching Level 5 is the end game for CSPs - and the power of GenAI will complement traditional AI capabilities to attain this goal. Traditional AI, which includes classical supervised, unsupervised machine learning, and reinforcement learning, has proven effective in handling structured telemetry data like time-series and tabular data. It’s been successful in use cases such as anomaly detection, delivering highly accurate outcomes. In contrast, GenAI excels in extracting knowledge from a wide range of unstructured data. Today, GenAI focuses primarily on natural language with multimodal enhancements and it’s best suited for complex but non-real time automation tasks in critical telecom environments; it acts as an AI assistant, augmenting the human decision-making process.

Now is the time for CSPs to take decisive action, and the good news is that the telecom industry is on time and ready to evaluate GenAI’s role in today’s Telco AI strategies.

Nokia was at the GenAI starting gate and has delivered measurable impacts already

In 2022, we partnered with a leading APAC CSP to develop a GenAI solution designed to achieve significant gains in network and service operations. This collaboration aimed to address challenges such as siloed domain knowledge, limited data sharing, high data analysis costs, and inefficient DevOps development cycles.

This partnership has delivered measurable impacts already, including an 80 percent reduction in knowledge acquisition time, a 72 percent increase in data analysis efficiency, and annual savings of 6.5M Euro in network operations. This success has created a blueprint for large-scale automation using GenAI, delving further into high-impact areas like automatic generation of network design configurations and optimization recommendations.

Guiding principles to help GenAI deliver successful outcomes in telecom

While the benefits of GenAI are exciting, it's important to address the risks and limitations. To achieve optimal results, we need to ensure robust and reliable responses. Low-quality responses can have negative effects, so it's crucial to infuse relevant knowledge into the foundational models and employ prompt engineering and fine-tuning techniques for accuracy and consistency.

We aim for knowledge augmentation in the telecom vertical to build Telco Gen AI-based practices. Our base models are internet-trained, providing a solid foundation. We navigate and build model farms by harnessing the power of both open and closed-source systems, ensuring optimal results for CSPs. Building telco-specific data vectorization, fine-tuning, and prompt engineering are critical components that remain a sharp focus for Nokia. We prioritize these aspects through key anchor projects, both internally and with our customers.

Domain and AI-modeling expertise drives Nokia’s GenAI strategy

At Nokia, we are fully committed to integrating GenAI into our AI and analytics solutions to power autonomous networks that ‘sense, think and act’. With our extensive knowledge across all network domains, proven AI leadership, and Bell Labs innovation, Nokia is armed with a federated knowledge source based on 30+ years of data accumulation. Leveraging this domain and AI-modeling expertise, we are connecting today's AI capabilities with GenAI’s future potential. Together, these synergistic forces will drive our customers closer to autonomous networks – and empower them to earn the title of AI-powered CSP.