The following blog was adapted from the transcripts of a video series. To view the videos, please click on the link at the bottom of the blog.
Why is AI Needed in Network Operations?
In the world of network operations, 1Finity (a Fujitsu company) offers 24/7/365 management and support for our customers’ networks. Our Network Operations Center (NOC) and Technical Assistance Center (TAC) covers multiple vendors, technologies, as well as broadband enterprise customers and all types of network operators. Anything can happen any time and we handle over 250 million alarms and more than 100,000 incidents each year with strict response Service Level Agreements (SLAs). Supporting network operations is fast paced, high volume, and high variety. In such an environment, Artificial Intelligence (AI) advanced algorithms can benefit operations and its customers tremendously, mainly to help achieve our objectives. We strive every day and every night to take care of our customers’ networks with a strong sense of urgency plus we constantly try to find ways to improve Mean Time To Repair (MTTR) resolution times, expedite root cause findings, and reduce escalations. Several applications of AI and Machine Learning (ML) help achieve this using generic AI, neural networking, or Generative AI.
AI also has the secret power of observability so it’s constantly learning, listening, and using AI to derive insights proactively for early warnings and finding trends. That’s extremely valuable for smart autonomous operations. AI, especially with its agentic framework, Retrieval-Augmented Generation (RAG) databases, as well as supervision techniques, allow you to get to those leaps of innovation that address true pain points and improve network operations. With AI, we can help our customers connect communities in a cost-efficient, secure, and scalable manner.
Use Cases and Applications for Using AI Within Network Operations
At 1Finity, we examine every aspect of network operations looking for areas where we can add meaningful improvement and intelligence to balance reactive and proactive approaches. So whether it’s automation, analytics, or now AI, there are several AI/ML use cases and applications underway in 1Finity’s network operations. In the case of 1Finity’s Managed NOC, our teams work 24/7 with multiple players involved, including fiber companies, power companies, Internet Service Providers (ISPs), multiple technologies, and Original Equipment Manufacturers (OEMs), which makes identifying root causes painstakingly complex and slow. To speed the process, we are leveraging neural network-based AI algorithms and have several explorations in the works to accelerate finding those root causes by improving alarm suppression, correlating alerts better, and ultimately improving MTTR resolution. AI-based assistance for workflow management, incident management, and deriving insights for early fault warnings are additional use cases.
Generative AI for TAC Support and Productivity in Network Operations
On the TAC support engineering side, there are tens and thousands of product documents and a troubleshooting knowledge base. We are using Generative AI-based Large Language Models (LLMs) to prompt trustworthy lookups and reduce escalation from days to seconds or minutes. We aim to see a step function increase in productivity by using AI agents to provide assistance in simpler functions of log parsing, code generation, query generation, and more. This way we can focus on elevating human intelligence for stronger reasoning capabilities and instilling greater confidence in customers.
What is 1Finity’s Approach to AI-Driven Network Operations?
1Finity is proud to be network experts and we have a lot of experience in both AI and network operations. In operations, we’ve developed a culture of automation and data analytics to drive tons of digital improvements and self-serve capabilities. So now, to be successful and effective in this AI universe, our approach is to combine the power of human intelligence as network experts and the magic of Machine Intelligence as algorithm experts. Several years ago, we transformed our entire operations infrastructure for our NOC and TAC customers from on-premises to the cloud, along with a whole host of applications on top of it. That enabled numerous automations, integrations, and visualizations for several functions such as incident management, service assurance, service activation, lifecycle management, among others. Now, with the advent of AI in operations, it helps organizations bring all of that big data from different technology stacks together and surface only the most relevant, actionable insights so they can apply the advanced algorithms to deliver more meaningful alerts and warnings. We co-innovate both with our internal Research and Development teams as well as our external partners to use AI/ML models and applications to improve response and resolution times, better the alarm level correlations for faster root cause, and then reduce escalation by looking up information and log parsing on Generative AI LLMs. Also, we can help our customers connect communities in a secure, advanced and scalable manner.
To learn more about 1Finity’s AI-powered network operations, watch the full video series.
