PINO G. DICORATO AI & NETWORK MANAGEMENT
established closed-loop network automation systems. With high-quality, network-trained data in place, AI will enhance existing network automation capabilities by providing intelligent decision-making and personalised service offerings, creating networks that are not just automated but truly intelligent. AI will uncover new patterns in the network through unsupervised learning and address many operational use cases that are not currently possible without significant human expertise. Three types of use cases will drive more efficient lifecycle management for network operators and provide immediate benefits: 1. Improving operational staff user experience using natural language 2. Minimising SLA risk through AI-driven troubleshooting 3. Meeting end-subscriber demand using intelligent planning IMPROVING OPERATIONAL STAFF USER EXPERIENCE WITH NATURAL LANGUAGE Enhancing knowledge sharing with generative AI, natural-language- powered documentation search and live network queries provides valuable insights into network-state conditions, quantitative measurements and network configuration attributes. These insights are delivered in a summarised and accessible format that enables operational staff to streamline network design planning and troubleshoot issues more efficiently. It also allows for faster diagnosis and resolutions, ultimately reducing downtime that could impact SLAs (Figure 1). Trained on embedded documentation, GenAI can enhance the overall user experience, reduce costs and improve network quality. Furthermore, embedded explainable AI mechanisms that provide clear references for the generated summary foster trust in AI-driven recommendations. MINIMISING SLA RISK THROUGH AI-DRIVEN TROUBLESHOOTING Developing troubleshooting tactics to resolve issues is onerous and time- consuming. It also requires a deep understanding of the optical network, meaning only a few experienced resources trained to address high-risk issues are available. AI-driven automation combined with agentic AI enables more staff to remediate optical network problems. The end result is more timely resolution to failures through simplified operational workflows. AI-driven automation and agentic deep reasoning enable contextual summaries from system KPIs, and can combine reviews of historical logs, correlating alarms and tabulating hypothesis tests against probable causes with high confidence (Figure 3). This
Figure 2 - AI-enhanced closed loop automation
AI-enabled automation helps ensure key stakeholders are notified and the most likely source of failure is addressed, limiting the time spent solving issues and improving overall operational efficiency. AI systems can also retain this knowledge for future occurrences. MEETING END-SUBSCRIBER DEMAND USING INTELLIGENT PLANNING Time to market is critical to business success. Several factors govern timely service delivery for network operators (e.g., ordering the necessary network equipment: shelves, transponders and add-drop shared-risk groups (SRGs) at each site location). This requires that both the capacity and the associated route constraints and terminal points on that equipment be configured to ensure the desired service is in alignment with end-subscriber SLAs. Monitoring infrastructure capacity consumption (optical channel utilisation) is essential to minimise the ordering process and deliver new network capacity where needed. The operational process is lengthy, given the resources needed to coordinate between optical planners and network engineering resources. This is where AI-driven automation delivers significant value. It provides digital workflows and AI/ML capabilities to analyse trends such as capacity utilisation and can forecast future requirements to support hands-free design, configuration and phased network rollouts (Figure 2). Automation provides a single view from the end-subscriber perspective, while incorporating much of the network- domain intelligence needed to plan optically, execute link-performance
feasibility, generate bills of materials, configure the infrastructure and service well in advance to ease supply-chain delivery. Combined, this enables operators to streamline decision-making and respond to changing conditions in real time. MAKING OPTICAL NETWORKS READY FOR WHAT COMES NEXT The future has never been more promising, or more demanding, than it is with the advancements in AI-driven automation. Networks need to evolve so they can perceive, reason, act, learn and adapt; by leveraging AI-enabled automation, they can shift from repetitive task execution to becoming self-healing. AI-enabled automation will include the capabilities needed to meet the technical demands of the next decade, all guided by human decision making.
Pino G. Dicorato, Director of Automation Solution & Technology, Nokia
www.opticalconnectionsnews.com
25
ISSUE 44 | Q3 2026
Made with FlippingBook Annual report maker