How Ericsson and AWS Are Pioneering AI-Powered Self-Healing Networks for the Future of 5G and Beyond

In the face of mounting complexity across global telecommunications infrastructure, the integration of artificial intelligence (AI) into network management has transitioned from experimental to essential. As 5G and emerging 6G systems demand lower latency, higher resilience, and extreme scalability, network providers must adapt to challenges that cannot be addressed through traditional human-centered operations alone. Recognizing this urgent need for transformation, telecom equipment giant Ericsson and cloud powerhouse Amazon Web Services (AWS) have announced a strategic collaboration to create AI-powered self-healing networks—a significant technological leap toward autonomous network operations.

This partnership, revealed in mid-2025, marks a notable milestone in the evolution of telecom infrastructure. It unites Ericsson’s telecom expertise and intent-based management platform with AWS’s advanced cloud-native AI tools and scalable compute infrastructure. At the core of this venture lies a shared commitment to automating the detection, diagnosis, and resolution of network anomalies in real time—capabilities that not only promise to revolutionize service quality but also significantly reduce operational expenditure (OPEX).

The traditional model of network maintenance is heavily reliant on human operators diagnosing alerts and dispatching corrective actions. This reactive approach is slow, error-prone, and ill-suited to manage the scale and speed of next-generation communications. Even the most well-staffed network operations centers (NOCs) struggle to meet the demands of real-time resolution across complex environments. Network downtime and degraded performance continue to plague telecom operators, often leading to financial penalties, churn, and reputational loss. The stakes are particularly high in industries such as autonomous transport, telemedicine, smart manufacturing, and public safety—sectors where milliseconds of latency or dropped connections could be critical.

To overcome these challenges, the Ericsson-AWS alliance aims to implement Agentic AI, a paradigm where autonomous software agents—rApps in Ericsson’s terminology—interpret intent-based commands and execute workflows without human involvement. These agents, when deployed across core, RAN (Radio Access Network), and transport domains, will close the loop on fault management by not just identifying issues but proactively fixing them. AWS’s AI stack, including foundational models from Amazon Bedrock and orchestration via AWS Lambda and Neptune, will serve as the intelligence backbone of this system, providing data-driven insights and orchestration agility at cloud scale.

Self-healing networks are not a novel aspiration within telecom circles. However, prior attempts were either rule-based, lacked contextual awareness, or failed to scale across vendor-neutral environments. The difference in this iteration lies in its use of generative and agentic AI, where models can generalize across varied use cases, simulate causal reasoning, and improve autonomously through reinforcement learning and telemetry feedback. This positions the Ericsson-AWS solution beyond legacy Self-Organizing Networks (SON) or robotic process automation (RPA), into the territory of truly autonomous networks as defined by TM Forum’s maturity levels (L4–L5).

From an industry perspective, this collaboration also signals an intensifying race among cloud providers and network equipment manufacturers to capture the future of telecom automation. AWS, already embedded in telecom via Wavelength and Telco Network Builder, is extending its footprint by integrating deeper into network operations. For Ericsson, this partnership reinforces its Cognitive Network Solutions (CNS) strategy and aligns with its broader mission to modernize telecom architectures with AI and intent-based systems.

In this blog post, we will explore the strategic scope and technical depth of the Ericsson-AWS partnership, unravel the core architecture of self-healing networks, examine business implications for telecom providers, and provide a future-oriented view of what autonomous networks might look like. We will also visualize how these AI systems operate, using diagrams and a comparison table to clarify the transformative potential of this initiative.

As the global telecom industry inches closer to a fully autonomous future, the fusion of Ericsson’s network knowledge with AWS’s AI capabilities might well define the blueprint for the next generation of resilient, adaptive, and intelligent communications infrastructure.

The concept of a self-healing network represents a fundamental shift in how telecommunications systems are managed, monitored, and maintained. In traditional network management frameworks, fault detection, diagnosis, and resolution are largely manual or rule-based processes that require constant human intervention. In contrast, a self-healing network leverages artificial intelligence and automation to autonomously identify and rectify problems in real time, thereby ensuring continuity of service, minimizing downtime, and significantly reducing operational overhead.

Core Definition and Functional Capabilities

At its core, a self-healing network is defined as a communication infrastructure that can detect anomalies, predict failures, initiate root-cause analysis, and apply corrective actions without human oversight. These networks are designed to be contextually aware, continuously learning from past disruptions and adapting to dynamic operating environments. Self-healing capabilities encompass both reactive and proactive strategies, with the ultimate goal of creating a zero-touch network operations model.

The primary functions of a self-healing network can be classified into four layers:

1. Anomaly Detection – Identifying deviations from normal behavior using telemetry and performance data.
2. Diagnosis and Root-Cause Analysis – Determining the underlying source of faults using AI inference and causal modeling.
3. Corrective Action Execution – Automatically implementing fixes such as re-routing traffic, reconfiguring devices, or restarting network functions.
4. Feedback Learning – Updating internal models and logic based on the outcomes of previous actions to improve future decisions.

Each layer operates in a continuous closed-loop fashion, often in milliseconds, thereby enabling networks to maintain optimal performance even during fluctuating conditions or unexpected disruptions.

From SON to Agentic AI: Evolution in Network Autonomy

The telecom industry has been working toward automation for over a decade, starting with Self-Organizing Networks (SON) in the 4G era. SON systems provided capabilities such as automatic neighbor relations (ANR), dynamic resource allocation, and cell outage compensation. However, these systems were largely deterministic and reactive, relying on predefined rules and thresholds that limited their ability to adapt to novel scenarios or scale effectively.

The rise of artificial intelligence, particularly machine learning and agentic AI, has redefined what is possible. Agentic AI introduces intelligent, autonomous agents—software entities that act independently to fulfill specified objectives. Ericsson refers to these as “rApps,” or Radio Network Applications, which operate on top of its Cognitive Network Platform. These agents not only observe network behavior but also interpret “intents”—declarative goals defined by operators—and autonomously execute the necessary steps to achieve them.

Unlike SON systems that react based on static rules, agentic AI leverages reinforcement learning, generative models, and graph-based reasoning to learn from complex environments. It dynamically builds causal maps of the network and can test “what-if” scenarios in real time, leading to better outcomes and fewer false positives.

Key AI Technologies Powering Self-Healing Capabilities

The Ericsson-AWS collaboration integrates a diverse set of AI technologies to achieve self-healing functionality:

• Generative AI: Used for natural-language intent translation and dynamic root-cause hypotheses generation. AWS Bedrock, with its foundation model hub, provides access to LLMs capable of understanding network intents and synthesizing solutions.
• Reinforcement Learning (RL): Powers agents that optimize decision-making through trial-and-error, balancing performance metrics such as latency, throughput, and energy consumption.
• Graph Neural Networks (GNNs): Enable multi-domain causal inference, where the topology and relationships between network components are considered in fault diagnosis.
• Agentic AI: Orchestrates autonomous decision flows using self-directed agents that monitor, reason, and act without requiring granular human input.
• Symbolic/Hybrid AI: Combines rule-based logic with statistical learning to ensure transparency, auditability, and regulatory compliance in decisions.

These technologies, when orchestrated via cloud-native platforms such as AWS Lambda, Neptune, and SageMaker, create an adaptive mesh of intelligence that sits across the network stack—from the radio access layer to the cloud edge.

Strategic Advantages and Industry Implications

The shift toward self-healing networks is not just a technical evolution but a strategic imperative. Operators that implement self-healing functionality stand to gain the following advantages:

• Reduced Downtime: Automated fault remediation can cut network outages by up to 80%, according to industry benchmarks.
• Lower Operational Costs: AI reduces the need for manual monitoring and troubleshooting, contributing to OPEX reductions of 30–50%.
• Enhanced User Experience: Improved network stability and responsiveness enhance quality of service (QoS) and reduce customer churn.
• Sustainable Operations: Energy-aware AI policies optimize power usage dynamically, aiding carbon neutrality goals.
• Service Agility: CSPs can launch, test, and scale new services—such as network slicing or ultra-reliable low-latency communication (URLLC)—with greater confidence.

The table below summarizes how different AI technologies map to specific functions and business outcomes within self-healing networks:

As the telecommunications industry transitions from static infrastructure to adaptive, intelligent systems, the concept of a self-healing network emerges as a cornerstone of this transformation. By uniting AI-driven autonomy with cloud-native orchestration, Ericsson and AWS aim to redefine how networks operate in real time.

The collaboration between Ericsson and Amazon Web Services (AWS) represents a landmark initiative in advancing the autonomy of telecom networks through a fusion of cloud infrastructure, artificial intelligence, and domain-specific automation. Positioned as a next-generation solution for achieving Level 4–5 automation as defined by the TM Forum’s Autonomous Networks framework, this partnership leverages the strengths of both entities: Ericsson’s deep-rooted telecommunications expertise and AWS’s mature, scalable AI infrastructure. In this section, we examine the core technical components of the collaboration, dissect its architecture, and evaluate how it facilitates the self-healing capabilities required in modern networks.

Strategic Integration Scope and Vision

The foundation of the Ericsson–AWS alliance lies in the mutual recognition of the need for dynamic, self-governing systems within the telecommunications landscape. The joint vision centers around three main objectives:

1. Operational Autonomy: Enable telecom operators to reduce manual intervention by replacing reactive operations with proactive, intent-based automation.
2. Cloud-Native Scalability: Provide elastic compute, storage, and AI model execution capabilities using AWS’s global infrastructure footprint.
3. AI-Driven Closed-Loop Control: Implement intelligent, automated workflows that can detect, analyze, and resolve issues in near real-time through agentic AI.

To support these goals, Ericsson has integrated its Cognitive Network Solutions (CNS) suite with AWS’s AI and orchestration services, allowing rApps (Radio Network Applications) to execute autonomously across cloud and on-premises environments.

Key Architectural Components

Ericsson’s Cognitive Network Platform and rApps

Ericsson’s CNS portfolio is the cornerstone of its automation ecosystem. At the heart of this architecture is the Cognitive Network Platform (CNP), a modular environment that enables the deployment and lifecycle management of intelligent applications—rApps—that support automation across the RAN, transport, and core domains.

Each rApp functions as a self-contained, AI-powered microservice designed to perform specific tasks such as:

• Intelligent load balancing
• Cell outage detection and compensation
• RAN parameter optimization
• Dynamic spectrum allocation
• Predictive maintenance scheduling

These rApps operate using declarative “intent-based” inputs, allowing operators to describe desired outcomes rather than micromanage how those outcomes are achieved. For example, an operator may issue an intent such as “maximize throughput in Zone B” or “minimize energy consumption during off-peak hours,” and the corresponding rApp autonomously determines how to fulfill that objective.

AWS AI and Cloud Infrastructure Stack

To provide the intelligence and scalability required for self-healing operations, AWS brings to the table an extensive portfolio of AI, ML, and orchestration services:

• Amazon Bedrock: Offers foundational model access from leading providers (including AWS Titan, Anthropic Claude, and Mistral) for generative tasks like natural language understanding, intent translation, and scenario simulation.
• Amazon Neptune: A graph database used to model complex network relationships and causality paths for root-cause analysis.
• AWS Lambda: Enables serverless execution of workflows, allowing rApps to respond rapidly to network events.
• Amazon SageMaker: Trains and deploys machine learning models that handle telemetry-based anomaly detection and prediction.
• AWS Wavelength & Outposts: Bring cloud-native services to the telecom edge, minimizing latency for critical decision loops.

This integration ensures that network intelligence is both scalable and proximate to where decisions must be made—whether in a centralized cloud or a distributed edge context.

How It Works: Data Flow and Decision Logic

The self-healing loop facilitated by Ericsson and AWS is designed to function autonomously and continuously. The data flow process can be described as follows:

1. Data Collection
   • Continuous telemetry from RAN, core, and transport layers is ingested.
   • Parameters include throughput, latency, jitter, energy usage, and error rates.
2. Anomaly Detection
   • AWS SageMaker models analyze streaming data for statistical and behavioral anomalies.
   • Real-time alerts are generated when performance deviates from expected norms.
3. Root-Cause Analysis
   • Graph-based AI models hosted on Amazon Neptune trace the anomaly’s source across domains.
   • Generative AI is used to simulate possible failure paths and recommend likely causes.
4. Intent Evaluation and Translation
   • An intent is either user-initiated or autonomously inferred based on SLA violations.
   • Amazon Bedrock models translate these intents into executable actions.
5. Autonomous Execution
   • rApps deploy pre-configured or dynamically generated solutions using AWS Lambda functions.
   • For example, an rApp might re-route traffic to unaffected nodes or recalibrate antenna tilt angles.
6. Continuous Learning
   • The outcomes are logged and fed back into AI models to refine future decision-making.
   • Reinforcement learning updates the policy models to better prioritize actions over time.

Use Case Examples: Proof of Concept and Early Deployments

Several proof-of-concept deployments and trials have been executed to demonstrate the real-world viability of the architecture:

• Orange Telecom Trial: Leveraging AWS’s AI stack and Ericsson’s CNS, Orange significantly reduced root-cause analysis time from several hours to under a minute in high-density urban RAN environments.
• Energy Optimization Pilot: A North American operator deployed rApps to reduce energy consumption during off-peak hours, achieving a 20% reduction without impacting service quality.
• Fault Recovery Simulation: In a controlled environment, the system autonomously identified a backhaul fiber cut and redirected mobile traffic using available microwave paths, all within 30 seconds.

Standards Alignment and Interoperability

A core pillar of this partnership is strict alignment with industry frameworks and open standards to ensure broad compatibility and operator control. Specifically:

• TM Forum Autonomous Networks Levels: The architecture targets Levels 4–5, where networks exhibit full autonomy and dynamic adaptation without human intervention.
• O-RAN Alliance: Ericsson’s rApps and SMO (Service Management and Orchestration) functions are designed to operate in multi-vendor O-RAN environments.
• ONAP Compliance: AWS provides orchestration hooks that are compatible with Open Network Automation Platform (ONAP), facilitating integration with existing telco workflows.

This adherence to open standards enhances the solution’s relevance to diverse CSPs and ensures that operator sovereignty is preserved even in a highly automated system.

The Ericsson–AWS collaboration offers more than a vision; it presents a robust, field-tested technical solution for enabling self-healing networks. By integrating Ericsson’s CNS rApps with AWS’s AI and orchestration stack, the partnership achieves a closed-loop operational paradigm that is scalable, flexible, and highly autonomous. Through a combination of intent-based programming, agentic AI, and cloud-native infrastructure, this solution addresses some of the most persistent pain points in telecom operations—namely downtime, complexity, and inefficiency.

The Ericsson–AWS collaboration to develop AI-driven self-healing networks represents a pivotal shift in the operational and strategic frameworks of global telecommunications. Beyond the technological innovation, this initiative has far-reaching implications for cost structure optimization, service quality enhancement, competitive differentiation, and market expansion. In this section, we assess the measurable business impacts for Communications Service Providers (CSPs), the broader strategic advantages for Ericsson and AWS, and the competitive and regulatory landscape surrounding this development.

Operational Cost Efficiency and OPEX Reduction

Telecom operators today face intensifying pressure to maintain service quality while reducing their cost base, especially in the capital-intensive 5G era. According to industry research, nearly 30–40% of a CSP’s total operating expenditure (OPEX) is allocated to network operations, including monitoring, diagnostics, manual fault resolution, and performance optimization.

The deployment of self-healing networks directly targets this cost segment by:

• Automating fault detection and resolution, thereby reducing the need for Level 1 and Level 2 technical support teams.
• Minimizing downtime and outages, which decreases compensation liabilities tied to Service Level Agreement (SLA) violations.
• Enhancing predictive maintenance, allowing operators to preempt equipment failures and avoid costly emergency interventions.

Initial field trials and simulations have shown promising metrics. For example, in deployments where rApps were used for energy optimization, operators reported up to 50% savings in site energy consumption during non-peak hours. Similarly, automated root-cause resolution workflows have demonstrated the potential to cut troubleshooting time by over 90%, translating into significant labor cost savings.

Quality of Service and Subscriber Retention

In an era where customer experience is a key differentiator, self-healing networks provide CSPs with a robust mechanism to ensure uninterrupted service delivery. AI-driven closed-loop control improves network availability, latency management, and traffic routing, particularly in congested urban areas or during large-scale events.

Benefits to customer experience include:

• Faster issue resolution, often before end users even notice a degradation.
• Reduced dropped calls and packet loss, enhancing voice and video quality.
• Dynamic optimization of bandwidth allocation, supporting services like 4K streaming, gaming, and AR/VR in real time.

These capabilities contribute to higher Net Promoter Scores (NPS), lower churn rates, and increased Average Revenue Per User (ARPU). For instance, by proactively maintaining service quality in high-density geographies, CSPs can better monetize premium 5G offerings and enterprise SLAs.

Strategic Advantages for Ericsson and AWS

From a vendor perspective, the partnership brings clear value differentiation:

• Ericsson gains a strong foothold in autonomous network leadership by positioning its Cognitive Network Solutions (CNS) as central to next-gen automation strategies. The company’s ability to deliver rApps that integrate seamlessly with AWS services extends its relevance in cloud-native telecom environments.
• AWS solidifies its presence in the telecom vertical, which has traditionally been cautious about adopting public cloud infrastructure. With telco-friendly services like AWS Wavelength, Telco Network Builder, and now AI integration through Amazon Bedrock and SageMaker, AWS is crafting a comprehensive, vertically integrated telecom AI platform.

Moreover, both companies benefit from being perceived as neutral enablers in a multi-vendor landscape—key in an industry that often resists vendor lock-in. Their alignment with open standards like TM Forum’s AN Levels and O-RAN Alliance principles enhances adoption viability across a wide array of global CSPs.

Competitive Landscape and Market Positioning

The move toward self-healing networks is not occurring in a vacuum. Competitors are actively advancing similar capabilities, and the telecom AI space is rapidly evolving:

• Nokia has invested in Digital Operations Center and the AVA cognitive services platform, which focuses on AI-driven service assurance and analytics.
• Huawei has developed its “Autonomous Driving Network” concept that similarly aligns with TM Forum Level 4–5 aspirations, albeit with limited global transparency due to geopolitical constraints.
• Juniper Networks and Cisco are integrating AI into their SD-WAN and service orchestration stacks, primarily aimed at enterprise network transformation.

Despite the crowded space, the Ericsson-AWS model differentiates itself through deep integration, modular rApps, and scalability across public and edge cloud, which few vendors can match in tandem.

Furthermore, by leveraging AWS’s hyperscaler capabilities, the platform offers global elasticity that is especially attractive to Tier 1 operators with transcontinental networks.

Risks, Challenges, and Mitigation Strategies

Despite the compelling benefits, the road to full-scale self-healing networks is not without hurdles:

1. Data Sovereignty and Security
   • Telecom data is sensitive, and many jurisdictions require local data processing.
   • The AWS–Ericsson framework mitigates this with hybrid models (e.g., AWS Outposts, Wavelength Zones), allowing edge deployments within compliance zones.
2. Model Trust and Explainability
   • AI-driven decision-making, especially when involving SLAs or billing, demands transparency.
   • Ericsson’s use of symbolic and hybrid AI models supports explainability, allowing CSPs to audit decisions.
3. Legacy Integration Complexity
   • Many CSPs operate brownfield networks with legacy equipment.
   • rApps are designed to operate across mixed environments, using APIs and adapters to interact with non-IP-based systems.
4. Organizational Readiness
   • Shifting to autonomous networks requires cultural transformation and new skill sets.
   • Ericsson and AWS have committed to co-developing training programs and operational playbooks to aid CSPs in the transition.

The integration of self-healing networks is far more than a technological enhancement—it is a strategic lever with the potential to reshape the economics, performance, and service models of global telecommunications. Through this collaboration, Ericsson and AWS are not merely building smarter systems but are catalyzing a new operating paradigm defined by autonomy, agility, and resilience.

For Communications Service Providers, the shift offers a path to sustainable growth in a margin-pressured environment. For the wider industry, it marks a turning point where cloud, AI, and telecom converge into a unified, intelligent infrastructure fabric.

The strategic collaboration between Ericsson and AWS is not merely a response to the immediate demands of modern telecommunications, but a foundational move toward redefining the long-term operational paradigm for global networks. As the telecommunications sector grapples with increasing complexity, mounting data volumes, and heightened customer expectations, the necessity for autonomy in network operations is more than a convenience—it is a competitive and operational imperative. This final section explores the future trajectory of self-healing networks, the emerging technologies that will reinforce their evolution, and the broader implications for industry transformation.

Towards Fully Autonomous Networks

The journey toward Level 5 autonomy, as outlined by the TM Forum’s Autonomous Networks framework, is incremental but inevitable. While many global operators are currently between Levels 2 and 3—where systems exhibit partial automation and decision support—the Ericsson-AWS framework is designed to facilitate the transition to Levels 4 and 5, where networks can make and execute decisions without human intervention.

In the next phase of development, we can expect:

• Increased domain fusion, where core, RAN, transport, and service layers are jointly optimized.
• Expanded use of intent-based orchestration, allowing human operators to define outcomes without prescribing exact processes.
• Greater reliance on digital twins—virtual models of the network that can simulate, test, and validate optimization strategies before real-world deployment.

These advancements will not only enable true end-to-end closed-loop operations but will also make it feasible to run networks with near-zero human intervention while maintaining high performance and resilience.

Emerging Enablers: Digital Twins, Edge AI, and Multi-Agent Systems

Three specific technological trends are set to further empower the development of self-healing networks:

1. Digital Twins
   • Digital replicas of entire networks will enable predictive modeling and stress testing of scenarios, facilitating proactive adjustments before faults occur.
   • Ericsson is already investing in network-level simulation platforms that integrate telemetry data to continuously train AI models.
2. Edge AI
   • AI workloads processed closer to the source of data will enhance latency-sensitive decision-making.
   • With AWS Wavelength and Outposts, edge-based execution of rApps will become increasingly feasible, especially for use cases like real-time handovers and ultra-reliable low-latency communications (URLLC).
3. Multi-Agent Systems (MAS)
   • A network of interacting AI agents, each with a specialized function (e.g., fault mitigation, energy optimization), will lead to more coordinated and efficient outcomes.
   • Agentic AI, as conceptualized in Ericsson’s rApp framework, is a foundational step toward scalable MAS in telecom networks.

Together, these technologies will converge to create intelligent, context-aware, and resilient networks that dynamically adjust to user needs, traffic fluctuations, and environmental disruptions.

Industry Implications and Ecosystem Transformation

The impact of this paradigm shift extends well beyond network engineering. The rise of self-healing networks will reshape the entire telecommunications ecosystem:

• For CSPs: Operators will be empowered to shift human capital from low-value monitoring tasks to high-value service innovation and customer engagement.
• For Vendors: The demand for AI-native, modular, and interoperable components will increase. Vendors that embrace openness, flexibility, and standardization will thrive.
• For Enterprises and Consumers: A more reliable network experience will underpin emerging applications in autonomous vehicles, smart factories, and immersive media.
• For Regulators: The autonomous nature of these systems will prompt new frameworks for AI governance, accountability, and cyber-resilience standards.

Moreover, the transformation will catalyze a services-first mindset, where the network becomes a programmable, adaptable fabric that supports dynamic enterprise workloads, customized SLAs, and agile product development.

Sustainability and the Future of Green Networks

One often overlooked but critical dimension of self-healing networks is their contribution to sustainability. With telecommunications infrastructure accounting for a substantial portion of global energy consumption, AI-powered optimizations can lead to:

• Reduced energy usage during off-peak hours through intelligent shutdown of unused network elements.
• Smart cooling and load balancing in data centers and base stations.
• Lower carbon emissions by minimizing unnecessary field interventions and transport logistics.

By embedding energy efficiency as a design principle, the Ericsson–AWS model contributes to achieving both environmental and economic objectives in tandem.

Conclusion

The Ericsson–AWS initiative to develop AI-powered self-healing networks marks a transformational step in the evolution of telecommunications. By merging Ericsson’s deep domain knowledge with AWS’s scalable AI and cloud infrastructure, the partnership delivers not just a product, but a new operating model for the industry. Self-healing capabilities rooted in agentic AI, intent-based orchestration, and closed-loop intelligence are no longer theoretical concepts—they are becoming operational realities.

As we look ahead, the trajectory points clearly toward fully autonomous, adaptive, and resilient networks that require minimal human oversight while maximizing performance, reliability, and sustainability. The implications span operational efficiency, customer experience, environmental impact, and competitive dynamics. For CSPs, the message is unequivocal: the future of telecom will be led by those who can master the art and science of intelligent, autonomous infrastructure.

With Ericsson and AWS setting a precedent, the race to operational autonomy is not only underway—it is accelerating. Those who fail to act risk being left behind in a landscape where agility, intelligence, and automation define success.

References

1. Ericsson – Cognitive Network Solutions
   https://www.ericsson.com/en/portfolio/networks/automation/cognitive-network-solutions
2. AWS – Amazon Bedrock
   https://aws.amazon.com/bedrock/
3. AWS – Telco Network Builder
   https://aws.amazon.com/telco-network-builder/
4. TM Forum – Autonomous Networks Framework
   https://www.tmforum.org/autonomous-networks/
5. Ericsson – Partnering with AWS for AI-Powered Self-Healing Networks
   https://www.ericsson.com/en/blog
6. AWS – Amazon Neptune
   https://aws.amazon.com/neptune/
7. AWS – AI/ML Solutions for Telecom
   https://aws.amazon.com/telecom/
8. Ericsson – rApps and SMO in O-RAN
   https://www.ericsson.com/en/reports-and-papers/white-papers/oran-and-rapps
9. TM Forum – AI and Autonomous Operations in Telecom
   https://inform.tmforum.org
10. AWS Wavelength – Ultra Low Latency for Edge Applications
    https://aws.amazon.com/wavelength/