Agentic AI in telecom OSS: Practical execution amid industry realities

The current state of agentic AI in telecom and CSP challenges

Artificial intelligence is recognized as a foundational driver for autonomous network operations, but the path to agentic AI in OSS is often blocked by common hurdles. As highlighted in Omdia’s 2026 report*, challenges include preparing telecom data, integrating AI agents with legacy OSS systems, and establishing strong governance control.

Let's examine several challenges in more detail:

• Data readiness: Legacy OSS architectures were not designed for AI-driven operations, resulting in fragmented, siloed, and inaccessible data. This impedes the ability to generate consistent inputs with context for actionable insights.
• Operational silos: Fragmented systems and processes create barriers to unified network automation and limit the scope of AI possibilities.
• Trust and governance: There is understandable skepticism regarding AI’s reliability, explainability, and compliance—especially when deployed to help operate critical communication networks.

These foundational barriers should be used to inform how agentic AI is designed and implemented—from data grounding and system integration to governance and control.

Blue Planet’s vision for agentic AI in telecom OSS

So, how does Blue Planet’s approach to enabling agentic AI solve for these common challenges?

Start with data.

Blue Planet, through its applications, solves this by bringing network and service data together into one unified view instead of scattered silos. Automated discovery and reconciliation keep data accurate and current, so it reflects the actual network states. With clean, connected data in place, agentic AI has the context it needs to make smarter decisions and take action–after all, You can’t AI what you can’t see.™

Next, operational silos.

Blue Planet removes operational silos by bringing inventory, orchestration, and assurance together on one shared platform across domains and vendors. Teams work from the same data and processes, with clear visibility across the entire network instead of disconnected tools. This unified view allows end to end automation, not just in isolated parts of the operation.

Last, governance.

Blue Planet builds trust by grounding AI in accurate context, based on information-rich data and proven operational processes. Clear guardrails and observability capture each AI agent’s actions, why they are acting, and keep them within operator control via robust access control and elicitation.

This lets CSPs use AI more confidently in network operations without sacrificing reliability or compliance.

Blue Planet is building a foundation for agentic AI rooted in open standards, information‑rich data, human oversight, and explainable outcomes. This approach leverages decades of OSS expertise and direct collaboration with leading CSPs.

Blue Planet focuses on:

• Bringing together data from all vendors and network domains into one clear view of operations.
• Turning that data into easy-to-use insights so teams can quickly see what’s happening and take action.
• Building advanced knowledge graphs capturing valuable network relationships to enrich understanding and decision‑making.
• Enabling continuous awareness of service state and underlying network health to support real‑time, proactive operations.
• Integrated APIs for automation that connect orchestration layers and business workflows to speed up processes and support closed-loop operations.

Blue Planet’s three-layer agentic AI framework

To move from vision to execution, it’s essential to have a structured framework that serves as both a roadmap and a toolkit for deploying agentic AI in telecom OSS environments. The Blue Planet agentic AI framework is designed to provide this structure, helping CSPs systematically overcome challenges and accelerate their journey towards autonomous network operations.

1. Agentic tooling
   Blue Planet's starting point for any AI initiative is data. The agentic tooling interfaces with OSS information across network domains and vendors, creating a unified operational view. This includes integrating inventory, topology, service, performance, routing, and assurance data. The goal is to ensure AI agents have access to comprehensive, contextual insights, which eliminate the need for prolonged integration projects. This data is exposed to AI agents from Blue Planet product APIs through the Model Context Protocol (MCP).
2. Agentic core
   The agentic core is the intelligence layer where agents are built, managed, and governed. Blue Planet AI Studio acts as the central hub for agent lifecycle management, orchestration, and operations. A secure LLM gateway allows CSPs to select models based on their capabilities and compliance requirements. This layer also embeds robust governance mechanisms, including access control, observability, and explainability. Every agentic process is transparent and auditable, ensuring that increased autonomy does not come at the expense of oversight.
3. Agentic channels
   AI delivers real impact when it drives measurable operational results. Blue Planet embeds AI directly within key OSS applications—inventory, orchestration, assurance—serving as built-in agentic channels. Integrating agentic capabilities throughout the portfolio enables higher productivity and drives meaningful OSS process transformation, while aligning automation outcomes with each CSP’s operational maturity and business priorities.

Figure 1. The three layers of the Blue Planet agentic AI framework.

The realities of delivering agentic AI in OSS

Putting agentic AI into action with Blue Planet doesn’t require years of prep work.

1. Identify key use cases
   Work collaboratively to document and evaluate OSS processes, identifying critical opportunities for agentic AI integration—whether to enhance employee productivity, supplement existing automation, or fundamentally redesign processes. Prioritize agentic initiatives according to impact and process complexity. This is a journey.
2. Select or build your agents
   Once use cases are defined and prioritized, the focus shifts to operationalizing the agents themselves. There will be no single platform or AI agent that meets every operational need, so autonomous networks demand an open, multi-agent ecosystem where agents can be sourced, built, and coordinated across vendors and domains. Blue Planet AI Studio is designed for this reality, enabling CSPs to use prebuilt agents, integrate partner or third-party agents, or build their own with low-code tools—while ensuring interoperability through open standards such as Agent-to-Agent (A2A) and the MCP. 
3. Access the right data
   Blue Planet helps streamline your systems by leveraging immediate access to rich, OSS-ready data. Our agentic AI framework is intentionally extensible, allowing agents to securely connect to third-party data sources and applications through open, standards-based interfaces. This makes it easier to get early agentic AI use cases up and running without waiting for large-scale data transformation projects. As adoption grows, teams can gradually modernize deeper data layers at a manageable pace—delivering near-term value while building a stronger foundation for the long term.
4. Integrate with operational channels
   AI is embedded natively within Blue Planet inventory, orchestration, and assurance, giving users direct access to AI agents’ capabilities. These agentic capabilities can also be invoked via other AI agents through A2A communication. Agent activity is continuously monitored, so teams stay in control and can refine automation as results improve over time.
5. Progressive implementation
   The framework lets CSPs start small and expand over time, scaling AI adoption as teams are ready and results are proven. Throughout execution, transparency and governance are non-negotiable. Agentic AI shall be designed to be observable, auditable, and aligned with CSPs’ operational policies.

The path forward for CSPs

As CSPs navigate the complexities of modernizing their networks, adopting an agentic AI framework provides a realistic path to transformation. By starting with practical use cases, leveraging accessible data, and integrating automation within existing operations, CSPs can take confident steps toward autonomous networks. Blue Planet’s agentic AI framework lays the groundwork for a more agile, intelligent network future.

*Omdia, Agentic AI for OSS/BSS: Opportunities and Challenges, Jan 2026.
Results are not an endorsement of Blue Planet. Any reliance on these results is at the third party’s own risk.