The surge of agentic OSS: Unlocking autonomous operations

The surge of agentic OSS: Unlocking autonomous operations

We know AI in telecom is big. But you might be surprised by its rapid growth. The global AI telecom market will rise from $4.4 billion in 2024 to more than $102 billion in 2033.^[1] Last year, 48% of communication services providers (CSPs) were piloting AI programs, while 41% were advancing to AI solution deployment.

We’re learning that the most impactful AI deployments unlock autonomous operations; they don’t recreate the silos and inefficiencies of legacy OSS. They should be purpose-built for telecom and designed to increase efficiency by removing operational silos.

AI can’t live in silos

Telecom has historically relied on siloed OSS platforms. These silos lead to incomplete network visibility, operational inefficiencies, and siloed data. And since AI thrives on structured, robust data, you cannot layer AI on top of isolated data or processes; you need a unified, holistic approach.^[2]

Today’s best solutions unify data across the network, domains, and vendors, avoiding the sins of the past.^[3] Removing silos is the only way AI has access to complete, accurate, and context-rich data to coordinate actions and deliver real business value. That’s because AI is only as smart as the data and architecture that power it.

In the words of industry expert, Joe Cumello: “You can’t AI what you can’t see.”^[4] Only by breaking down silos and providing AI with comprehensive visibility can your organization realize the full potential of automation.

On the journey to delivering service innovation and operational efficiency, legacy inventory systems create major roadblocks. CSPs are looking to unify inventory to access a real-time, end-to-end view across multi-vendor networks.^[5] We’ve recently seen Lumen do this successfully, allowing their teams to isolate faults in minutes, reduce new service introduction processes by weeks, and achieve radical operating efficiencies.^[6],[7]

Federated and structured data is critical

One of the keys to enabling autonomous networks – like the one Lumen is transforming to – is cloud-native orchestration to provide integration across both physical and virtual infrastructure. This is only possible with true cloud-native tools, however. Wrapping functions in a Kubernetes wrapper won’t provide the flexibility inherent to cloud-native environments. When combined with open approaches to ensure simplified operations across vendors and domains, cloud native OSS future-proofs the network, reduces costs, and accelerates service delivery.

With these requirements in mind, CSPs are turning to a modern breed of OSS partners which are more advanced in their AI and cloud native approaches to automate the complexity of planning, deploying, and managing multi-vendor equipment. As expected, what we’re seeing is that by unifying inventory, orchestration, and service assurance capabilities, onboarding times are drastically reduced, configuration errors eliminated, and data is streamlined. ^[8]

One example of this is what we’ve seen with Blue Planet. Their AI-friendly approach to building autonomous networks allows agents to access deep telecom-specific context for every decision. In the context of bandwidth management in municipalities, a CSP can use a knowledge graph to rebalance network resources proactively during crowded city events. The combination of real-time data from 5G small cells and IoT devices helps CSPs avert bottlenecks while strengthening network resilience. ^[9]

Coordinated, intent-driven agents: The agentic AI difference

Unifying and cleansing network data in real time ensures AI agents have up-to-date, reliable insights across domains. Empowered by this structured data, AI agents can access network context and trigger automated responses when anomalies, such as unusual traffic patterns or emerging faults, are detected. The agents can reroute traffic or initiate preemptive maintenance.

Foundational to autonomous networks, agentic AI and digital twin technologies are key to producing higher levels of operational self-management. Moving to autonomous networks could generate more than $800 million in annual value for CSPs, with the greatest benefit seen as operators reach levels 4 and 5 on the TM Forum’s Autonomous Network model.^[10] These stages require intent-driven, collaborative AI agents for real-time orchestration and optimization.

Telecom-native agents are on the market today to help interpret high-level business intent, such as prioritizing connectivity for emergency services or ensuring continuous compliance, and translating those objectives into coordinated actions across complex, multi-vendor networks.

With these agents onboard in testing environments around the world, network operations are being simplified by empowering both technical and non-technical staff to create and deploy agents for use cases like fiber fault detection or predictive maintenance, accelerating automation without deep programming skills. Supervisor agents are ensuring that all agent actions are aligned with business priorities and policies, while user-friendly chatbot interfaces make it easy to trigger and monitor automation.

Agent-to-agent (A2A) automation frameworks enable CSPs to coordinate critical tasks, such as rolling out software patches across global footprints or dynamically optimizing resources during major events, while maintaining oversight and flexibility. Agents continuously share and analyze operational data, autonomously resolving issues or escalating for human input when needed. By embedding digital twin insights and proactive responses, a premier agentic AI solution gives CSPs the tools to scale, adapt, and maximize value as they progress to next-generation autonomous networks.

Cloud support and integration

The percentage of CSPs who say they’ve moved at least 50% of OSS/BSS workloads to the public cloud has more than doubled since 2021.^[11]

CSPs accelerate innovation and automate complex workflows by pairing telecom AI solutions expertise with cloud providers’ scalable AI infrastructure, maintaining data security and compliance. They can deploy solutions on-premises or in any cloud.

These CSPs need their agentic framework to integrate deeply and seamlessly with public cloud services like AWS Bedrock, as demonstrated with a 5G slicing solution at Digital Transformation World.^[12]

An effective telecom AI platform needs to support deployment across AWS, Google Cloud, and Azure natively and integrate with public cloud AI services. This enables CSPs to automate advanced scenarios such as dynamic, real-time network slicing. For instance, a CSP can use a telecom AI platform on Google Cloud to adjust network slices automatically during city events with surging connectivity demands, ensuring reliable and optimized service quality for thousands of concurrent users.^[13]

OSS AI value – today

Starting from scratch to enable these new AI use cases can be daunting. Savvy OSS vendors are providing pre-built AI use cases such as fault prediction, traffic forecasting, and service management, drawing on deep telecom expertise, giving CSPs access to the AI “easy” button.  A drag-and-drop, LLM-agnostic platform lets users deploy or develop agents rapidly, using any major AI model or custom toolkit.

The journey to autonomous networks is more than a technology upgrade; it transforms how you deliver value to compete in a crowded market. A premier OSS partner streamlines data and simplifies operations across multiple vendors and domains. The modern OSS builds in agentic intelligence with telecom-native AI agents that communicate and act across the cloud-native ecosystem. The modern OSS replaces rigid workflows, breaks down silos, speeds innovation, and delivers on the promise of the future-proof techco.