- A new Fierce Network Research report delves into the interesting ways that operators are using AI in their network field operations
- Rakuten Mobile is using AI in conjunction with photos to check the work of field technicians
- Rakuten is also using AI to detect "sleeper cells" in its network
Traditionally, field operations have been a labor-intensive part of telecom networks with technicians having to use a lot of their own judgment, and that judgment might vary based on experience.
Takai Eddine Kennouche, AI lead at the office of the CTO with Viavi, spoke with Fierce Network Research for a new report entitled “AI-Driven Network Testing.” Kennouche said, “Technicians have to travel with their trucks to service locations to perform significant manual labor and make on-site decisions. This human element, while essential, introduces variability and an observability blind spot.”
But now, some operators are using cutting-edge AI technologies to improve telecom network field operations.
How AI is revolutionizing telecom field operations
For the Fierce Network Research report, we spoke with Anshul Bhatt, chief product officer in the OSS Business Unit at Rakuten Symphony, who said one interesting way that Rakuten has been using AI in the field is via the use of photos to verify that work is done properly.
Although photos have been used for site audits before the advent of AI, in the past the photos would have to be manually checked by a person in a central office. And if that person wasn’t immediately available, then the technician might leave the site and move on to another job.
Rakuten Mobile’s AI-driven approach to network optimization
At Rakuten, when an engineer completes his field work, he can take a photo, and then Rakuten-based AI technology immediately checks to see if the installation was done correctly, or not. If something is amiss — such as weatherproofing missing on pipes, for example — the technician can make the appropriate fixes before leaving the site for his next job.
Rakuten is also using AI in the field to find “sleeping cells,” in its mobile networks. These cells can appear to be well-functioning, but they are actually delivering poor customer experiences, sometimes for days on end.
Rishi Davda, head of the Network Intelligence Department at Rakuten Mobile, said in a webinar that in his experience working for operators, only a very tiny percentage of an operator’s cells will go into sleeper mode per day. But it’s a consistent problem, that must be continuously addressed. And historically, it’s been a very time-consuming problem for engineers because they had to monitor alerts and reports to identify sleeper cells. It was akin to looking for a needle in a haystack.
By using AI, it has become much easier. Davda said at first Rakuten Mobile was working with a limited data set. But as automation software becomes trained on the problems of sleeper cells, and the data set gets larger, the accuracy and detection become easier. Once a sleeper cell has been identified, the cell can be reset, solving the problem.
Davda said, “Our day-to-day life isn’t only about improvements. We want to always advance and improve, to make life more easy for individuals or the engineers or whomever is working on the problems, to have better
Of course, Rakuten Mobile and Rakuten Symphony are ahead of the game in terms of AI. They’re working with a greenfield mobile network, which had the foresight to put everything in one massive data lake.
For the complete report about AI in telecom networks, click here. Other topics in the report include the use of digital twins in networks, catching network anomalies with AI, and how AI might affect the telecom workforce.