- Verizon is leveraging a full spectrum of AI, including predictive AI, computer vision and generative AI
- Predictive AI and computer vision help Verizon optimize 5G node placement and improve customer experience
- Meanwhile, generative AI, through tools like Network Genie, can reduce cognitive load for technicians
While generative artificial intelligence (GenAI) dominates much of the conversation in tech circles today, we shouldn’t completely forget the old AI days, if you ask President of Verizon Global Services Shankar Arumugavelu.
In an interview with Fierce Network, Arumugavelu said that Verizon is focused on harnessing all forms of AI—ranging from predictive AI to computer vision and generative AI—as it works toward a long-term goal of network autonomy.
“We’re using AI across the entire spectrum of all functions within the company,” he said.
A complete AI arsenal
While GenAI has gotten most of the attention since the launch of popular bots like ChatGPT, Arumugavelu stressed that other types of AI are still invaluable across the network lifecycle.
“It's a combination of everything. There is still tons of applicability for good old predictive AI,” he said.
Predictive AI remains foundational in Verizon’s network planning. For example, computer vision technology helps Verizon optimize the placement of 5G nodes in dense urban areas. By analyzing 3D GIS maps and applying AI, Verizon determines the best locations to maximize coverage with minimal deployment for a more cost-efficient and optimized network build-out.
On the operational side, Verizon uses predictive AI to improve the customer experience by analyzing over 180 network features and narrowing them down to the 26 most critical factors that influence network quality. This enables the company to compute a “quality of experience” score for each customer and proactively address potential issues on the network.
But the hype isn’t for nothing. As part of a broader AI kit, GenAI can do some things that the other tools just can’t. It's a great customer service tool, for one. And it can help reduce cognitive loads for technicians with the help of natural language processing (NLP).
Through a system known as the “Network Genie,” Verizon technicians can ask a question and quickly get AI-generated answers, making the process of network maintenance faster and more efficient, noted Arumugavelu. Basically, GenAI allows operators to talk to their data, and thereby their networks, with more ease than before.
Verizon builds toward autonomy
While GenAI is the shiny new toy, a blend of AI types creates a more powerful approach to advancing network intelligence, Arumugavelu argued.
“We believe in reinventing end-to-end workflows using a combination of predictive, prescriptive and generative AI," he said.
Verizon’s ultimate goal is an autonomous network, one that can run, heal and optimize itself without constant human intervention. Arumugavelu, who also serves on the board of TM Forum, referenced the industry-standard definition of an autonomous network that moves through stages from fully assisted operations to self-healing, fully autonomous systems.
However, Verizon is taking a measured approach. “We’re not fully autonomous yet,” he said. Currently, the Verizon network operates with human oversight, though many functions are already automated or semi-automated. This human-in-the-loop approach allows the company to manage risk while advancing toward a future where certain network functions are entirely autonomous.
“We have not publicly put a timeline on when we will be fully autonomous,” Arumugavelu concluded, but added that Verizon continues to integrate more autonomous functions into its operations.