A new era of core network automation
Ericsson’s vision for intelligent core automation reflects a clear shift from today’s partially automated, human-dependent networks to truly autonomous, AI-native, cross-domain, and intent-driven networks that keep network behavior continuously aligned with business intents and outcomes. In this future state, the core network is no longer governed by static, reactive rules; instead, AI-powered, agentic closed loops interpret high-level intents, observe live network conditions, and take autonomous actions across domains to enhance performance, resilience, and overall experience at scale. The journey will span multiple years, but the direction is already defined: an autonomous, AI-native core in which self-configuration, self-optimization, self-healing, and self-protection are inherent properties of the network fabric rather than externally added features.
The power of intent-driven automation
Central to this vision is intent-driven core automation, where communications service providers (CSPs) articulate business priorities in a declarative form, such as target service levels, energy-performance objectives, or enterprise service level agreement (SLA) obligations, instead of specifying detailed, imperative configurations. AI-native capabilities, including machine learning, agentic AI, and advanced reasoning, then translate those priorities into actionable insights and anticipatory, closed-loop decisions. These AI-driven loops continuously reconcile the as-designed and the as-lived network state, dynamically managing multiple, and at times competing, objectives around cost, performance, energy usage, and resilience.
Intelligent lifecycle management and capacity planning
Realizing this vision in the core starts with higher performance in deployment, lifecycle management, and capacity planning. Network automation and AI are applied from day-zero design through day-N operation to streamline deployment, shrink integration timelines, and ensure that capacity is always right-sized to actual demand. Intelligent, AI-driven orchestration makes this possible, turning capacity engineering into a continuous, data-driven process rather than a periodic exercise as is currently the case. In parallel, capacity planning focuses on smart auditing of network dimensioning to ensure that deployed systems are accurately dimensioned. This leads to increased resiliency, avoiding network incidents caused by incorrect dimensioning and potential misconfigurations, while ensuring optimized use of hardware resources.
When the network heals itself
At the same time, the vision depicts a highly resilient and secure core network, enabled by self-healing by design, effective change management and proactive threat mitigation, all supported by end-to-end observability. Autonomous network operations introduce service observability, anomaly detection, and predictive analytics so that degradations and failures are detected before they manifest as incidents. AI agents correlate key performance indicators, logs, and topology across domains to identify root causes and to recommend or execute remediation, improving customer experience and reducing churn. Coupled with Ericsson’s broader work on trustworthy, secure, AI-driven networks, this positions the core network to withstand both operational faults and evolving cyber threats while maintaining regulatory-grade reliability. The network function anomaly and core root cause explainer are concrete steps on this path, with the core network playing a key role in network resilience—our customers’ top priority.
From best effort to guaranteed
Intelligent core automation is also a key enabler of premium experience, unlocking value in differentiated connectivity and fraud prevention. Differentiated connectivity shifts the business model from best effort to an assured, tiered, AI-managed promised experience where specific performance attributes—such as latency, throughput, reliability, and prioritization—are guaranteed and monetized. Intent-driven, utility-aware automation allows premium intents to be fulfilled, with SLA observability and prediction ensuring performance is provable and auditable. In parallel, AI-powered analytics on signaling, usage, and behavior patterns in the core provide the basis for proactive fraud prevention and protection of high-value services, further strengthening trust in premium offerings. Differentiated connectivity, powered by AI in the core network, thus becomes a key opportunity to enable service differentiation.
Building the autonomous core
Taken together, these capabilities form an intelligent core automation blueprint for autonomous network realization. It relies on AI-native cApps running on the EIAP as the core domain controller—the vehicle to implement autonomous networks delivering an open, secure, scalable platform that meets operators’ performance and innovation requirements. Leveraging intent-based operations and agentic closed loops, the core evolves into a set of autonomous domains that are increasingly self-governing yet remain explainable and controllable by humans.
One platform, endless possibilities
A key differentiator of the EIAP is its openness and simplicity for developers. By relying on clear, standardized interfaces (R1) and a comprehensive developer toolkit, including software development kits, Ericsson Intelligent Controller (EIC) enables seamless onboarding of cApps through a one-stop environment that provides access to curated, real-time network data—spanning both RAN and Core, as well as the underlying infrastructure, enriched with topology awareness. In addition, machine learning operations (MLOps) capabilities bring a full set of model lifecycle services into EIC. These capabilities support data, model, and code engineering for cApps, decouple the model lifecycle from the application lifecycle, and enable controlled deployment including declarative and progressive approaches, continuous monitoring, and automated model retraining under drift detection, thereby providing a robust MLOps backbone for EIAPs AI driven, closed loop automation use cases. EIC relies on Ericsson Network Manager (ENM) and Ericsson Orchestrator (EO) as the primary data sources for network data and sends controlled and coordinated actions for closed loop operations based on decisions made by the cApps.
Collect once, use everywhere
This data foundation is further strengthened by the evolution of ENM through capabilities such as Ericsson Stream Processing and Enrichment (ESPE), which introduces native streaming support into the network management layer. ESPE enables data to be collected once and efficiently made available to the EIC and cApp developers, eliminating fragmentation and accelerating innovation cyclesUnlike traditional approaches, ESPE continuously collects, processes, enriches, and curates network data in real time, exposing it through scalable mechanisms such as Kafka streaming and replication. At the same time, it ensures that northbound consumers can receive anonymized and curated datasets, simplify data governance while making high-quality data readily usable for AI-driven use cases. This provides a streamlined and trusted data pipeline—critical for enabling autonomous operations over time.
Focus on innovation, not integration
This environment exposes consistent, cross-domain data and vertical insights—combining metrics across network layers, services, and infrastructure—allowing cApps to operate with full contextual awareness. This abstracts away the complexity of data access, security, and network actuation, allowing developers to focus entirely on innovation rather than integration. As a result, cApp development becomes faster, safer, and more scalable, fostering a vibrant ecosystem, building on the momentum already seen with rApps.
Test, iterate, scale
Critically, EIAP enables a try and fail fast innovation model. Unlike traditional approaches—where new functionality must go through lengthy operations support systems (OSS) integration, testing, verification, and long-term system maintenance—cApps can be rapidly onboarded, tested, and iterated in a cost-efficient way. Developers can also discover and reuse existing cApps, accelerating development cycles and enabling faster realization of new ideas.
This capability not only drives the introduction of new autonomous self x operations, but also enables the modernization and consolidation of legacy OSS-based automation solutions. Many CSPs today rely on fragmented, custom-built systems to support automation tasks—often difficult to maintain, dependent on scarce expertise, and lacking modern security, scalability, and lifecycle management.
Reducing technical debt, raising the bar
With EIAP, these legacy capabilities can be progressively re-implemented and streamlined as cApps, moving from siloed, code-heavy environments into a unified, cloud-native platform. This allows operators to preserve existing automation intent while simplifying and standardizing execution, reducing technical debt, improving maintainability, and strengthening security. As a result, EIAP becomes not only a platform for innovation but also the target environment for rationalizing and future-proofing exiting landscapes, making these capabilities accessible even to developers without deep network domain experts.
At the same time, Ericsson cApp preparation is reinforcing EIAP by making core domain knowledge more readily available, for example through model learning capabilities.
From technology to business value
For customers, this translates into tangible business outcomes: avoidance of penalties due to incidents, churn reduction based on improved user experience, and lower operational expenditure through reduced manual effort and faster incident resolution. Additionally, higher SLA adherence and accelerated time-to-market for differentiated 5G and enterprise offerings are realized. By turning the core into an AI-native, intent-driven engine, CSPs can both unlock new revenue streams from premium connectivity and systematically reduce the risks and costs associated with misconfigurations and fraud, ultimately strengthening competitiveness and profitability in demanding markets.
Read more:
