Dynamic 5G services, made possible by AI and intent-based automation

The emergence of 5G networks opens a new frontier for connectivity, enabling advanced use cases that require ultra-low-latency, enhanced mobile broadband, and the Internet of Things (IoT) at scale. However, behind the promise of this hyper-connected future lies an equally critical challenge: managing the complexity of 5G networks. With dynamic demands on bandwidth, latency, and reliability, traditional static configurations and manual operations are no longer sufficient.

Enter intent-based service management, a pioneering approach that combines AI and automation with service orchestration and assurance. With intent-based service management, network resources in a 5G network scale and adjust dynamically according to real-time demand, based on the business intent, the 5G network resources are orchestrated dynamically and adjusted just as they are for cloud computing, storage, and networking and streaming services. ​

Together, Ericsson and Google Cloud are pushing the boundaries of this technology, exploring  technology that not only simplifies operations but redefines how telecommunications companies can deliver on the promise of 5G.

The power of intent-based autonomous operations in 5G

Intent-based networking has entered the spotlight as Communications Service Providers (CSPs) recognize that they must master substantial network complexity with sophisticated automation if they want to create, deliver, and sustain services, especially in cloud-native network environments like standalone 5G.​ Automation of networks and related operations cannot evolve and mature to full closed-loop autonomy without intent-based implementations.

Take for example a hospital that needs dedicated 5G connectivity, provided by a 5G network slice, in order to support emergency communications with ultra-low latency. The required configuration in such a scenario should not be a static configuration;  in order to provide optimal service experience it should evolve with the network conditions and the demand. The network layer needs to understand the intent behind every service request  e.g. “provide ultra-low-latency video services” — and translate it into actionable configurations across the network, which then are dynamically updated based on real-time factors such as changes in traffic conditions, underlying network status, and so on.

With Ericsson’s Service Orchestration and Assurance and Google Cloud’s Vertex AI, this process becomes far easier. The solution takes high-level intents, expressed in natural language, and maps them to precise technical requirements like bandwidth, latency, and throughput; then orchestrates the creation of a tailored network slice in real time. The quality of network-slice-based service provisioned is assured via the intelligent closed loop automation mechanism.

TMF standardization for intent-based operations​

To integrate different solution components with one another without expensive and fragile integration solutions, you need a standards-based interface and model.

The Telemanagement Forum (TMF) has specified APIs and a model with common intents to enable this sort of integration. TMF provides intent models and interfaces at the business, service, and resource layers, which derive from CSP-defined business objectives and follow the Autonomous Networks Reference Architecture.​ Ericsson and Google Cloud’s approach is fully compliant to these standards and intent models.

The Role of AI and automation

At the heart of this solution lies AI-driven closed-loop automation. Once a service is instantiated, AI agents continuously monitor its performance against defined KPIs such as quality of service (QoS). Going back to our hospital example, if latency in the hospital’s network slice exceeds the acceptable thresholds, the system automatically identifies the issue and proposes corrective actions, such as reallocating RAN resources or Core Network.

This closed-loop system isn’t just reactive — it’s predictive. By analyzing data from the network in real time, it can anticipate issues and resolve them before they impact users. For telcos, this means not just operational efficiency but also delivering a superior  quality of service.

Ericsson’s closed-loop system leverages Google Cloud’s AI agents for the intent translation, proposal and evaluation agents for slice assurance and remediation. The AI agents reason on how to best achieve a goal based on inputs and tools at their disposal, leveraging models, tools, and orchestration.

In addition to Gemini-based AI agents, Ericsson Service Orchestration and Assurance uses Vertex AI to implement predictive AI services for the remaining closed-loop agents (such as assurance, measurement, and actuation). 

Why Ericsson Service Orchestration and Assurance and Google Cloud

Ericsson Service Orchestration and Assurance is a multi-domain, multi-technology platform that enables cross-domain orchestration across access, transport, and core networks. It addresses CSPs’ critical need for simplification of open and programmable networks by automating the service lifecycle, ultimately boosting revenues and profitability.

The platform provides an open, standards-based architecture that fosters innovation and multi-vendor integration in the telecom ecosystem. It allows CSPs to design services with ease across both network and IT environments, reducing time-to-market and enabling collaboration across diverse teams and use cases with fewer manual errors. 

Furthermore, intent-driven orchestration accelerates the testing and launch of new services in multi-vendor environments, potentially increasing the number of services that can be launched. SLA commitments are enhanced through closed-loop automation, improving customer satisfaction and operational efficiency. Designed to be vendor-, technology-, and service-agnostic, Ericsson’s platform empowers CSPs to implement multi-domain service orchestration effectively, enabling service expansion and revenue growth.

To support this sophisticated framework, Google Cloud provides a foundation to meet the demands of modern telecom networks:

• Scalability for dynamic networks: As 5G networks expand, Google Cloud’s elastic compute power helps ensure even the most resource-intensive processes, such as intent translation and real-time assurance, run seamlessly without performance bottlenecks.

• AI-first infrastructure: Leveraging Vertex AI, intelligent agents translate intents into actionable configurations while optimizing network performance. With AI tools like BigQuery ML, telcos can move beyond basic analytics to actionable insights, enabling smarter decision-making.

• Vertex Model Garden for open innovation: Pre-trained models and open-source LLMs in Vertex AI Model Garden empower telcos to innovate and customize highly specialized 5G service management solutions.

• Hardware acceleration: Google Cloud’s Tensor Processing Units (TPUs) and Nvidia GPUs provide the computational power necessary to efficiently train and deploy complex AI models for intent-based service management.

• Hybrid and edge capabilities: 5G networks demand hybrid setups across cloud and edge environments. Google Cloud’s edge-to-cloud orchestration helps ensure operations, whether optimizing central data centers or delivering ultra-low-latency services at the edge.

• Security and compliance: Operating in a highly regulated environment, telcos benefit from Google Cloud’s secure, reliable infrastructure tailored to meet the stringent compliance requirements of telecom workloads.

• Gemini for advanced reasoning: Gemini’s multimodal models enhance intent interpretation by processing natural language and network telemetry data. This enables nuanced and precise intent translation, leading to more effective automated actions, higher efficiency, and improved reliability of 5G service management.

Together, Ericsson Service Orchestration and Assurance and Google Cloud are exploring technology to meet these challenges together, combining automation, intelligence, and scalability to empower CSPs to deliver innovative, efficient, and profitable 5G services.

Real-world impact: A hospital use case

Let’s return to the hospital scenario. A request for a network slice is received from hospital users through the system’s intent translator: "Provide video service for patient doctor interactions having ultra low latency.​"

The solution springs into action:

1. Translating intent: The request is converted into technical configurations for the RAN and core network.

2. Creating the slice: Using orchestration tools, the system allocates resources and configures the network slice with Quality of Service (QoS) goals across the 5G network.

3. Assuring performance: With closed-loop automation, AI agents continuously monitor quality of service elements such as latency or packet loss. If a deviation that violates the intent occurs or is forecasted to occur, corrective actions are proposed and executed automatically.

In this scenario, the hospital’s needs are met dynamically, helping to ensure critical operations proceed without disruption.

Beyond efficiency: Redefining 5G Innovation

Intent-based service management enables telcos to innovate at scale. By automating complex tasks, telcos can focus on creating new services, entering new markets, and unlocking revenue streams. For instance, they can offer differentiated slices tailored to industries like gaming, manufacturing, and healthcare, creating a competitive edge in a crowded market.

Moreover, the integration of custom gen AI models like Gemini helps these solutions be tailored to telcos’ unique needs. These LLMs, trained on telecom-specific data, enable intelligent automation that feels intuitive and responsive.

Ericsson and Google Cloud’s collaboration showcases the potential of intent-based service management. By combining Ericsson’s technical expertise with Google Cloud’s AI infrastructure, telcos can navigate the complexities of 5G with confidence.

As 5G adoption accelerates, solutions like this will become the cornerstone of telecom operations, enabling seamless, dynamic networks that deliver on the promise of connectivity.

The future of 5G isn’t just about faster speeds or lower latency — it’s about intelligent, intent-driven networks that adapt to the world’s needs in real time. And together, Ericsson and Google Cloud are making that future a reality.