Dominic Jainy stands at the forefront of the intersection between artificial intelligence and telecommunications infrastructure. With a career dedicated to the nuances of machine learning and the practical deployment of enterprise software, he brings a unique perspective to how “agentic” systems are reshaping the BSS/OSS landscape. In this discussion, we explore the shift toward autonomous service orchestration, the critical role of standardized APIs, and how predictive models are becoming the new standard for customer retention and network evolution. We delve into the mechanics of multi-layer control planes and the necessity of maintaining vendor-neutral AI architectures in an increasingly complex digital ecosystem.
How does implementing agentic AI across customer care and service configuration fundamentally change the daily operations for telecommunications providers?
The introduction of the Customer Care Agent is a significant shift because it doesn’t just display data; it proactively surfaces relevant service information to resolve inquiries on the fly. By determining the next best action automatically, we see service teams achieving much faster and more consistent results while significantly cutting down the onboarding time for new staff who no longer need to memorize every legacy process. On the technical side, the Service Configuration Agent handles the heavy lifting of complex system integrations, which traditionally required a rare breed of specialists. This automation reduces manual effort and the dependence on that scarce specialist knowledge when new services or network integrations go live. Ultimately, it allows the human workforce to move away from repetitive data entry and focus on higher-level problem solving and strategy.
With the industry moving toward open standards, why is having 47 TM Forum Open API certifications such a pivotal milestone for this new version of the suite?
Having 47 certifications, 15 of which were just added since the previous version, provides a bedrock of interoperability that is virtually unmatched in the independent BSS/OSS market. These standards have already been proven across the installed bases of seven different customers, demonstrating that the framework is ready for mission-critical tasks rather than just experimental labs. When you have this many open APIs, the AI agents can communicate across different systems without the friction of proprietary silos or custom-coded bridges. This openness ensures that as carriers scale their operations, they aren’t hitting a brick wall of integration costs or technical debt. It also facilitates the “composable” architecture that modern operators demand to remain agile in a fast-paced market.
The latest Enterprise Product Catalogue includes a “competitor-aware” design feature; how does this influence the way companies now approach product innovation and market response?
The system is now designed to ingest details of rival offers directly, which allows the agentic package creation tool to inform real-time decision-making within the Enterprise Product Catalogue. This significantly shortens the loop between a competitor making a move and a provider launching a counter-proposition to protect their market share. We are also seeing the benefits of the TM Forum ODA Component Certification for Product Catalogue Management, which validates that the architecture is mature and standardized. Furthermore, the Dynamic Catalogue Workspace introduces a grid-based environment that lets teams edit complex pricing and packaging structures at scale, applying those edits directly. This makes the whole process feel more like a surgical strike than a slow administrative task, allowing for rapid experimentation with new service bundles.
Managing multiple AI agents requires a sophisticated oversight mechanism; could you explain the significance of the three-layer control plane in maintaining operational security?
Governance is handled through a very deliberate structure starting with the AI Management Centre, which centralizes the oversight of all active models to ensure they remain within operational bounds. Below that, the MCP Server uses a Progressive Discovery architecture to provide secure, authenticated access to the Open APIs, keeping the data flow protected from unauthorized internal or external shifts. The third layer, known as the Agent2Agent Server, is perhaps the most critical because it coordinates the multi-step workflows that jump between different agents and external systems. This setup allows operators to adopt the vendor’s own models or bring in-house or third-party AI at any point in the stack. Because the architecture is deliberately model-independent, it prevents CSPs from being locked into opaque operational tooling that they cannot control or audit.
Machine learning is often cited as a tool for retention, but how do these predictive churn models translate into measurable outcomes within the Promotions Engine?
The predictive churn models work by running machine-learning algorithms that score each customer based on their departure risk, identifying the subtle “tells” that suggest a user might be looking for a new provider. These scores are fed directly into the Promotions Engine, which triggers targeted retention campaigns tailored to the individual’s specific needs and history. In recent customer pilots, these models proved highly effective at identifying at-risk accounts before the customer even considered reaching out to a competitor. By automating this entire flow, the system moves from reactive firefighting to a proactive strategy that saves high-value accounts through timely, relevant offers. It effectively turns data that was previously just sitting in a database into a primary driver of revenue stability.
As the industry navigates the shift to 5G Standalone, how does the new native Nchf interface assist providers who are still managing legacy infrastructure?
The native Nchf interface acts as a vital bridge because it allows REST-based 5G charging and older Diameter-based charging to coexist within a single, unified deployment. This means hybrid networks can keep running smoothly without a “rip and replace” scenario that would disrupt current revenue streams or frustrate existing customers. It provides a clear, low-risk migration path to 5G Standalone by handling the technical complexities of charging across different generations of network technology. By integrating this directly into the BSS/OSS suite, providers can focus on rolling out new 5G features while the backend handles the heavy lifting of monetization across the entire network footprint. This balance of innovation and stability is exactly what large-scale operators need to stay competitive during long-term transitions.
What is your forecast for the evolution of agentic AI in the BSS/OSS sector?
I expect that we will see a rapid transition where AI agents move from being simple assistants to becoming the primary orchestrators of the entire service lifecycle. We are already seeing the groundwork with 47 certified APIs and model-independent architectures that allow for a “plug-and-play” approach to intelligence. In the coming years, the friction between network configuration and commercial product launches will nearly vanish as these agents learn to sync market moves with technical capability in real-time. This will lead to a “zero-touch” environment where the system self-corrects and optimizes without needing constant human intervention. Ultimately, the providers who embrace this open, agent-driven framework will be the ones who can pivot their entire business strategy in days rather than months.