Less Buzz, More Bandwidth: The Near-Term Reality of AI in Telecom

AI Insights Series #1

Over the past two years, AI has exploded onto the scene in most industries. As a sector that lives and breathes technology by nature, it’s no surprise that telecom has particularly embraced large language models (LLMs) as tools to be integrated into products and operations. 

Nearly two years ago, our team at 7Sigma developed IRIS, an AI tool that helps users pinpoint issues and take action on important network insights generated by our NOC360 application while helping our customers cut down their subscriber call and customer service representative (CSR) training times. As one of the first companies in the industry to leverage AI with a team that is excited about monitoring and testing new tools, we’ve gained valuable insights into the most effective ways to utilize this technology in the near term. 

Reining in the hype of AI in telecom

As companies become increasingly excited about harnessing AI, the impulse to create flashy tools has increased, but many lack the ability to produce valuable insights. Most of us have probably experienced this firsthand, noticing that applications we use daily roll out AI features that don’t feel particularly useful or effective. Telecom products are no exception; the AI-generated predictions and summaries they offer can sometimes be helpful, but often clutter dashboards without making meaningful contributions to efficient workflows or improved operations. 

Excitement about AI can also lead telecom vendors to push its adoption on a large scale with the hopes of achieving total automation to ultimately replace humans. However, the reality is that, at least for now, LLMs don’t have these capabilities - and shouldn’t be expected to. Even today’s best models fall short in a variety of areas, making it both impossible and irresponsible to depend on them fully. 

Shifting focus to augmentation and actionable insights

So, how should AI be utilized in the telecom industry? Looking to the near future, companies must produce AI products not simply to be a part of the hype and appear to be embracing the newest technologies, but to have a meaningful and positive impact on internet service providers and carriers by helping reduce costs, improve operational efficiency, decrease subscriber churn, and the like. 

Tools that support humans, not replace them, should be the focus of these efforts. Telecom professionals’ individual field experience, intuition, and ability to make personal connections make them an invaluable part of business operations today. Meanwhile, many AI models' greatest strengths lie in their abilities to recognize correlations in large datasets that are otherwise difficult for humans to identify and provide actionable suggestions for addressing network issues. Working in tandem, these unique skillsets make it so AI helps humans work smarter, adopt more efficient workflows, and make better-informed decisions. 

Getting the best of both worlds with IRIS

Our AI layer, IRIS, is a primary example of how actionable insights from AI and human knowledge work together to solve network problems. Beyond summarizing a report on the status of PONs or predicting what could be wrong with an ONT, the tool suggests next steps for team members to take, such as dispatching a technician to check the power supply to the ONT or investigating whether damage to fiber optic infrastructure is the cause of an outage. IRIS is particularly helpful for CSRs, providing recommendations to relay to subscribers and helping decide whether issues should be escalated. These insights produce measurable benefits for ISPs, including up to 70% shorter subscriber call times, significantly reduced CSR training times, and more efficient workflows that shave minutes off the issue troubleshooting and resolution process. 

Tools like IRIS create a win-win situation for telecom. Strategically leveraging AI to support professionals means that workflows become more efficient, organizations have greater capacity, employees keep their jobs, costs are reduced, and subscribers are satisfied with their service. 

Interested in learning more about NOC360 and IRIS? Schedule a call with our team.