Google Cloud and the rise of the agentic telco

Technological shifts have always been the catalysts for opportunity in telecommunications. Today, we’re entering the most significant transition yet: the agentic AI era. To date, AI has focused primarily on efficiency — bolting on tools to perform manual tasks faster. Now, we’re moving beyond siloed tools toward a fully agentic ecosystem that can automate entire workflows.

Google Cloud provides a unique value for the telco industry by connecting our global AI infrastructure to the entire mobile ecosystem — spanning Android, Pixel, and iOS alike. We unite the entire value chain— bridging the gap between the data center and the device to make the network invisible and the experience effortless.

As we head into Mobile World Congress (MWC) Barcelona, here is how we’re helping telecommunications companies adopt agentic AI and wire intelligence into the core of their business.

Data: The foundation of agentic AI

For an AI agent to do more than just answer questions — to actually take accurate and complex action across systems — it needs to understand the business as well as a human expert. Most telcos, however, are navigating "data swamps”: massive amounts of fragmented information trapped in disconnected silos. 

Google Cloud is helping operators turn this "noise" into a strategic asset in three key ways:

• Automated data lifecycle: We're leveraging BigQuery’s latest evolution to automate the heavy lifting of data preparation. By integrating the latest Gemini models directly into the data lifecycle, we’ve automated the complex work of cleansing and tagging massive amounts of network telemetry, turning raw data into AI-ready information for telcos in a fraction of the time.

• Creating a digital map: To give AI the context it needs to act, Cloud Spanner Graph serves as a planetary-scale digital twin — a live, temporal map that unifies every relationship across the network, so that that AI can sense exactly where a problem is and how it affects the system.

• Building the pipeline: Through a new partnership with DigitalRoute, we’re launching reusable data pipelines. These act as high-speed filtration systems, transforming chaotic network signals into a single source of truth.

When these three areas work together, agents are grounded in reality, allowing them to sense a problem, reason through a solution, and act — often before a telco customer notices an issue.

Orchestrating the digital core

Today’s telcos must meet the needs of changing consumer behavior and deliver experiences that offer immediate, frictionless value. Yet, they are often hindered by a fragmented maze of disconnected legacy software. We’re helping solve this by providing both the unified data and AI foundation — via BigQuery, Spanner, and Vertex AI — as well as the agentic platform via Gemini Enterprise.  

Rather than relying on people to manually sync data between departments, Gemini Enterprise acts as a "reasoning engine." It weaves an agentic fabric across telco business systems (BSS/OSS), allowing intelligent agents to bridge the gap between technical network events and actual customer outcomes. As a result, we enable our telco customers and their consumers to find, understand, and autonomously act on information across a complex enterprise environment.

Engineering the autonomous network

To bridge the gap between 5G investment and revenue, we’re helping telcos reach Level 4 and 5 autonomy: a network that identifies, diagnoses, and fixes its own problems without human intervention. At MWC, we’re demonstrating how our Autonomous Network Operations framework has evolved into a production reality, powered by three core product innovations: 

• The network digital twin: We’ve evolved the digital twin from a static map into a temporal graph powered by Cloud Spanner Graph. This allows agents to "see" the network's live state and query historical data — seeing how the network looked hours or days ago—to perform instant root-cause analysis. Working with NetAI, we’re helping MasOrange use these models to resolve incidents with machine-speed accuracy.

• Unified graph data layer: We’re breaking down the silos between operational and analytical data. By leveraging Spanner Graph for digital twins and federated graph analytics through BigQuery, telcos can interoperate between real-time updates and deep historical analysis without slow, manual data transfers. To accelerate this, we’re open-sourcing our telco data pipeline and models on GitHub, allowing CSPs to implement industry-standard ontologies immediately.

• Predictive healing with GNNs: By training Graph Neural Networks (GNNs) in Vertex AI, the network moves from monitoring to predicting. It mathematically tracks how failures propagate, resolving them before they impact subscribers. 

Vodafone is using our framework to allow their network to fix its own connections, while our work with Deutsche Telekom on MINDR uses multiple agents to resolve issues across different network domains simultaneously.

Outside of our own product updates, we’re working with the broader ecosystem to integrate specialized capabilities into our framework. We’re launching agents with Future Connections, including an Autonomous Data Steward to automate data governance, and autonomous network agents for voice core and OSS networks. One NZ is currently deploying these network agents, moving beyond monitoring to active execution — like independently re-routing traffic or resetting network settings to restore call quality the moment a drop is detected. And finally, in partnership with Nokia, we’re integrating its Network-as-Code (NaC) platform with our AI stack, allowing engineers to use natural language to instantly prioritize network resources for critical services like emergency response or remote healthcare.

Building the proactive loyalty engine

In an industry that often struggles with low customer satisfaction, the root cause is usually a break-fix model: something breaks, the customer gets frustrated, and then they have to call to fix it. 

We’re helping operators break this cycle by building a proactive loyalty engine. Instead of waiting for a complaint, this engine uses AI to identify potential issues — like a Wi-Fi issue or a billing error — and fixes them automatically.

We’re collaborating with Amdocs to launch an agentic telco contact center, fusing Google Cloud’s Gemini Enterprise for Customer Experience with Amdocs’ industry-specific intelligence. Central to this solution is Amdocs’ Cognitive Core, the generative AI foundation of its agentic operating system (aOS), which runs on Google Cloud to provide the high-performance infrastructure needed for scale.

By connecting conversational AI directly to the core systems that run the business (like billing and network management), an agent can identify an outage, reset a connection, and apply a credit to a bill — handling the entire resolution before the customer makes a complaint.  

Accelerating revenue growth

Growing Average Revenue Per User (ARPU) remains a challenge as operators face an imbalance between heavy network expenditures and flat returns from commodity data plans. To bridge this gap, our cross-Alphabet initiatives help telcos move beyond selling megabytes to monetizing their network as a secure identity platform. 

A prime example of this is our work with Deutsche Telekom to create the CAMARA Number Verification 2.0 API. By leveraging the GSMA Open Gateway standards, we’re helping transform the network into a source of truth for digital security. This allows for easy network authentication, where a user is verified instantly and securely in the background — eliminating the friction and the fraud risk of traditional SMS passcodes. Instead of just charging for a text message, telcos can now offer Identity-as-a-Service (IDaaS) to millions of developers, turning the network into a value-creating engine. 

The Future: One Google, every hand

The agentic AI era represents the greatest opportunity since the launch of the smartphone. By integrating Gemini across the full stack — from the core network to the phone in a subscriber's hand —we enable a continuous thread of intelligence. Together with our partners, we’re making the network invisible and the experience effortless for the everyday.

To see the agentic AI era in action, visit the Google Cloud Booth (Hall 2, #2H40) from March 2–5 at Mobile World Congress. We’re also hosting a dedicated agentic customer journey in the booth, with showcases dedicated to each of these pillars, and live demonstrations of multi-agent systems (MAS) like autonomous service resolution, digital twins, and physical AI robots. To learn how the latest AI advancements enable new consumer experiences for the Android ecosystem, check out Android Avenue between Hall 2 and 3 or read here.