Cerillion announced the general availability of version 26.2 of its composable BSS/OSS suite, embedding a set of “agentic AI” features aimed at automating customer‑care, service configuration, product design and churn prediction for communications service providers (CSPs). The upgrade arrives as telecom operators accelerate 5G rollouts and look to AI‑driven processes to cut costs and improve time‑to‑market.
Agentic AI at the core of the release
The 26.2 update introduces four AI‑powered agents built on Cerillion’s AI Management Centre, MCP Server and Agent2Agent Server. Unlike generative models that create new content, these agents act as autonomous assistants that retrieve, process and act on data within the BSS/OSS stack. The architecture is model‑agnostic, allowing CSPs to plug in in‑house models, third‑party services or a mix of both while retaining centralized governance.
Customer Care Agent – faster ticket resolution
Designed for call‑center operators, the Customer Care Agent surfaces relevant subscriber and service details, suggests next‑best actions, and can auto‑populate response fields. According to Cerillion, the tool reduces average handling time by up to 20% in pilot deployments, while also shortening the onboarding curve for new agents because the AI supplies contextual prompts.
Service Configuration Agent – automating network onboarding
The Service Configuration Agent expands automation to complex provisioning steps, such as orchestrating virtual network functions (VNFs) and updating inventory records. By codifying specialist knowledge into reusable workflows, the agent aims to cut manual configuration effort, a critical need as CSPs integrate 5G standalone (SA) charging and hybrid legacy networks.
Competitor‑aware Package Creation – data‑driven product design
Within the Enterprise Product Catalogue, the new Competitor‑aware Package Creation agent ingests publicly available offer data and feeds it into the Enterprise Product Catalogue’s design engine. Product managers can therefore generate differentiated bundles faster, responding to market shifts without exhaustive manual benchmarking.
Predictive Churn – early warning for retention teams
Leveraging machine‑learning models trained on usage, billing and interaction histories, the Predictive Churn agent produces risk scores that feed directly into Cerillion’s Promotions Engine. Early tests reported a 15% lift in retention conversion when targeted offers were triggered based on the AI’s scores.
Open standards and API leadership
Cerillion emphasizes its compliance with TM Forum Open APIs. The suite now holds TM Forum ODA Component Certification for Product Catalog Management (TMFC001) and has added 15 additional Real‑World Open API certifications since version 26.1, bringing the total to 47 across seven customers—a claim the company backs with TM Forum’s publicly available certification registry.
Why the upgrade matters for CSPs
Telecom operators face mounting pressure to monetize 5G, reduce churn, and deliver omnichannel experiences. Agentic AI promises to embed decision‑making directly into operational workflows, reducing reliance on separate analytics platforms. Moreover, the open, composable architecture means CSPs can evolve their AI stack without vendor lock‑in, a point highlighted by Frost & Sullivan analyst Vikrant Gandhi.
Competitive context
Major BSS/OSS vendors such as Amdocs, Nokia and Ericsson have also introduced AI modules, but many are packaged as monolithic suites or tied to proprietary cloud services. Cerillion’s claim to a “model‑independent” layer differentiates it, positioning the company as a potential integrator for CSPs that already run heterogeneous AI workloads on platforms like Microsoft Azure or Google Cloud.
Potential limitations and risk factors
Agentic AI effectiveness hinges on data quality and the robustness of underlying models. CSPs with fragmented data lakes may see limited accuracy until they invest in data‑cleansing pipelines. Additionally, the automation of configuration tasks raises governance concerns; mis‑routed changes could impact network stability if proper audit trails are not enforced.
Broader AI ecosystem implications
The 26.2 release illustrates a shift from “AI‑as‑insight” toward “AI‑as‑action” in enterprise telecom software. By exposing agents through open APIs, Cerillion enables integration with AI cloud platforms (e.g., AWS SageMaker, Azure Machine Learning) and LLM‑based assistants that could later augment the existing agents with natural‑language capabilities. The approach also underscores the growing importance of AI orchestration layers—similar to what AI infrastructure providers like NVIDIA and Habana are building for edge workloads.
Market Landscape
Adoption of AI across telecom BSS/OSS is accelerating. IDC projects worldwide AI spending will exceed $500 billion by 2024, with telecoms accounting for a growing share as operators seek to automate network operations and customer engagement. Gartner predicts that by 2027, 70% of CSPs will have deployed AI‑driven BSS functions, up from 42% in 2023. Within this context, open‑API‑first strategies are gaining traction because they reduce integration costs and allow vendors to interoperate with cloud‑native AI services. Cerillion’s expanded certification portfolio aligns with TM Forum’s roadmap emphasizing composable, API‑driven ecosystems for 5G and beyond.
Top Insights
- Agentic AI embeds autonomous decision‑making directly into BSS/OSS workflows, moving beyond traditional analytics dashboards.
- Cerillion’s open, model‑independent stack lets CSPs mix proprietary and third‑party AI models while retaining centralized governance.
- Predictive churn scores integrated with the Promotions Engine have shown a measurable lift in retention campaign effectiveness.
- Certification with TM Forum’s Open API standards positions Cerillion as a leading integrator in the increasingly composable telecom stack.
- Successful deployment still depends on high‑quality data and robust change‑management processes to mitigate operational risk.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
