Verizon just made a move that few expected to see this early in the telecom industry.
The American carrier put Claude Code, Anthropic’s agentic coding tool, in the hands of 33,000 technology employees, essentially turning that entire team into software developers overnight.
And it doesn’t stop there.
While all of this is happening internally, AI agents are already running on Verizon’s network around the clock, detecting failures, resolving issues, and making configuration changes autonomously — all without a human needing to press a single button.
The goal is ambitious: reach Level 4 network automation, the highest tier on the scale defined by the TM Forum, where decisions and actions happen in closed loops without direct human intervention in execution.
But don’t worry, humans aren’t being left behind in this story. 😉
What’s changing is their role within this new model — and understanding that difference is essential to grasping what Verizon is really building.
In this article, you’ll see how it all connects, from the democratization of software development to the carrier’s vision for Open RAN and 6G.
Receive the best innovation content in your email.
All the news, tips, trends, and resources you're looking for, delivered to your inbox.
By subscribing to the newsletter, you agree to receive communications from Método Viral. We are committed to always protecting and respecting your privacy.
Claude Code in the hands of 33,000 people: what this actually changes
Distributing an agentic coding tool to tens of thousands of employees isn’t a small decision, and Verizon clearly didn’t take it lightly. The beginning was modest: in January, fewer than 500 employees were using Anthropic’s tool to develop customer-facing applications. That’s when Yago Tenorio, Verizon’s CTO, realized the potential went far beyond that. In just six weeks, access was opened up to the entire technology team.
Claude Code works differently from what most people picture when they hear about a coding assistant. It doesn’t just suggest lines of code like a glorified autocomplete. It acts, executes, tests, and iterates autonomously within a controlled environment, functioning as an agent that can pick up a task and carry it through to completion. Tenorio summed up this shift nicely when he told Light Reading that about a decade ago, there was a dream among carriers that everyone could be a software developer. It was tough back then, but now, according to him, you can imagine it, and Claude can code it for you.
The most immediate impact of this decision is the democratization of software development across the company. Network engineers, systems analysts, infrastructure specialists, and other professionals who have never written a single line of code in their lives can now build scripts, automate repetitive tasks, and create internal applications using natural language to instruct Claude Code. Tenorio described the usage as covering practically everything: engineering, modeling, capacity planning, capital allocation, forecasting, and automation. The core idea, in his words, is to turn the team’s expertise into software and remove complexity by developing an abstraction layer on top of the operating system.
But what happens when tens of thousands of people start developing software on their own? Doesn’t that sound a little chaotic? Tenorio acknowledges the concern but calls it a good problem to have. According to him, suddenly you turn every person into a parallel IT department — and that’s a much nicer problem than dealing with the complexity and bottlenecks of traditional software development. He also pointed out that people have been very disciplined and that the carrier maintains clear, well-defined guardrails to keep that freedom from turning into a mess.
It’s worth noting that Verizon isn’t married to a single tool. The company also uses Gemini, Google’s coding solution, and remains open to other options that may emerge in the market. That flexibility shows a mature strategy focused on results rather than blind loyalty to any one vendor.
From Dan Schulman’s strategy to the AI-native company concept
All of this transformation gained momentum after Dan Schulman took the helm of the carrier last October. The CEO is on a clear mission: turn Verizon into not just an AI-first company, but something even deeper — an AI-native company, where artificial intelligence isn’t an add-on but part of the fundamental fabric of how everything works.
The massive adoption of Claude Code across the technical team is one of the most visible faces of this shift. Tenorio explained that since Schulman’s arrival, the company has been revolving around delighting the customer, and that the autonomous network effort fits perfectly into this philosophy of putting the consumer at the center of everything. In his words, this story contributes directly to building the new Verizon.
AI agents running on the network: how autonomous automation works
While Claude Code operates on the development side, another layer of AI agents is already working directly on Verizon’s network infrastructure — and this part of the story is perhaps even more impressive. These agents continuously monitor network behavior, identify anomalies, and when they detect something wrong, they can spin up sub-agents to pinpoint exactly which domain or network element is causing the problem. The practical result is that the carrier can identify and resolve failures in under two minutes.
The numbers are striking. Last year, Verizon carried out more than 70 million network configuration changes autonomously through its closed-loop automation platforms. This concept is central to understanding the level of network automation the company is pursuing. In traditional systems, an alarm is generated, an analyst is notified, they analyze the problem, consult a manual, execute the commands, and document the action. That entire cycle can take anywhere from minutes to hours. In a true closed loop, the AI agent detects, analyzes, decides, acts, and documents — all in seconds, without any human needing to be present during execution.
Tenorio stated that the carrier is deploying these automated loops slowly but surely, using more and more tokens at each stage. The target is to reach Level 4 network automation in the near future, spanning all domains and use cases as defined by the TM Forum. Currently, the company is working to move past Levels 2 and 3 — classified as conditional autonomy — and advance toward the highest tier on the scale.
The two pieces that make all of this possible
Two elements underpin this entire structure. The first is what Verizon calls a common data layer, which brings together near-real-time telemetry on network performance. This system provides such a granular view of the customer experience that it can break down what’s happening floor by floor inside buildings. The second is a common set of skills, tools, and agents. Tenorio explained that the idea is to transfer human knowledge and expertise into skills and tools that agents can use, allowing those learnings to be shared across the entire organization.
With these two pieces in place, deploying agents across different network domains becomes straightforward, because the agents connect to the network through purpose-built software development. And that’s exactly where the massive use of Claude Code ties directly into the carrier’s ambition for autonomous networks.
Humans are still in charge
Reaching Level 4 automation means that closed-loop decisions and actions previously handled by people are now carried out by systems. But that doesn’t mean humans will no longer be needed. Quite the opposite. According to Tenorio, even when agents make decisions autonomously, there will always be a human ensuring the outcome is correct.
He was emphatic in defining the new role of people in this scenario. Humans are the owners of the agents — they define the guardrails, the scopes, and the frameworks, and most importantly, they are the source that trains the agents with skills. In his words, there will always be humans, but their function changes. Instead of carrying out mechanical, repetitive tasks, they shift to creating, teaching, and training agents. The autonomous network removes humans from the execution loop and creates a different loop — a much better one.
This vision dismantles the simplistic narrative that AI-based network automation eliminates jobs. What changes isn’t the need for people, but the nature of the work. Engineers stop spending their shifts responding to alarms and start working on policy definition, curating AI models, and anticipating structural problems that agents still can’t identify on their own.
Open RAN, 6G, and the strategic role of AI in this equation
There’s a common thread connecting all of this activity to Verizon’s long-term positioning, and it runs straight through Open RAN. At first glance, linking AI agents and network automation to the need for Open RAN in future 6G architecture might seem like a stretch. But Tenorio connects the dots clearly: open interfaces make it possible to deploy autonomous networks and interconnect intelligent systems on top of them. Regardless of which language models end up being used, what truly matters is having standardized interfaces to support agentic applications that can be deployed on the radio network.
This becomes critical for what Verizon calls context-aware AI agents, which the carrier envisions running on future mobile networks. The idea is that agents or systems of agents are permanently aware of their surrounding environment and receiving data without the user needing to manually feed the system.
Tenorio gave a pretty personal and curious example. He built his own running coach using Anthropic’s technology and said he’s never trained so well — though he still has to manually send the data from each run. In the future, he imagines running with some kind of wearable that is aware of his pace and running technique, offering real-time guidance. That type of application, which constantly feeds data back to the network, would push uplink capacity and require far more dynamic traffic management — something only possible with advanced network automation.
While Open RAN isn’t required for network automation itself, it is a non-negotiable architectural foundation for Verizon’s 6G vision. Currently, nearly half of the carrier’s radio access network is Open RAN-based, supplied by Samsung running on Intel x86 hardware, while the rest, supplied by Ericsson, is not. Tenorio stressed the importance of Open RAN maturing across multiple vendors before 6G arrives, so that generic off-the-shelf hardware can also support this next generation. He even mentioned that the processing power of Intel’s Granite Rapids should be sufficient to fuel the early years of 6G.
What this move reveals about the future of telecom
Verizon’s moves aren’t happening in a vacuum. They reflect a broader transformation across the entire global telecom industry, where carriers are shifting from being pure connectivity providers to positioning themselves as technology companies with advanced AI agent and automation capabilities. This repositioning has a direct impact on competitiveness, operational costs, and the agility to launch new services. Companies that master this transition earlier will have a significant structural advantage.
The fact that Verizon chose Anthropic’s Claude Code as a central tool in this journey also says a lot about the maturity that AI agent tools have reached. Two years ago, distributing an agentic coding tool to tens of thousands of employees would have been considered a high-risk experiment. Today, it’s a strategic business decision with clear productivity and return metrics. That leap in perception is a powerful indicator of where we are on the AI adoption curve in large-scale enterprise environments.
What becomes crystal clear when you look at all of this is that AI-powered network automation is no longer a future promise for Verizon. It’s an expanding operational reality, with agents already working autonomously, tens of thousands of employees equipped with agentic tools, and a strategic vision that connects the present to the future of Open RAN and 6G. The pace of this transformation is accelerating, and the next few years will show just how much this bet will redefine industry standards. 🚀
