An AI-driven operational model for the age of differentiated Telco experiences

The CSP Imperative: reinvent for enterprise growth

The telecom landscape is undergoing a profound shift. With mobile data traffic growth slowing down and consumer markets saturating, communications service providers (CSPs) are pivoting to the enterprise sector as a strategic growth engine. This shift demands far more than connectivity. CSPs must now deliver tailored, premium, SLA-backed experiences across 5G, fixed broadband, private networks, and digital services. To win in this market, CSPs need to:

• Guarantee differentiated quality of experience (QoE) to demanding enterprise verticals and deliver real-time, on-demand tailored connectivity services.
• Assure and monetize 5G slices, private networks, and application-level SLAs.
• Demonstrate and report on service performance to their customers in real-time and allow customers to make configuration changes as they would operate their own private networks.
• Operate with predictive intelligence and AI-driven autonomy.

This transformation is both complex and urgent, but the benefits are higher win rates, higher revenues from enterprise customers and overall reduced opex because of operational automation. All this requires a unified operational foundation powered by data, automation, and AI.

A new operating model for the enterprise-focused CSP

To deliver and monetize SLA-based enterprise connectivity services, CSPs must update their commercial and pricing models to enable transactional business and update their Business Support Systems (BSS) accordingly. Focusing on their operating model, they must implement 3 key evolutions: from network to service centric, from reactive to predictive, and from technical to business intent. Elaborating on this:

1. Service-centric operations

This operating model enables CSPs to model, monitor, and manage service performance end-to-end, from onboarding through to activation and usage. The service-centric process involves:

• A focus on generating real-time service quality insights and impact analysis for all digital services for every enterprise customer.
• Monitoring of application-level SLA, from the infra all the way up to the application layer as this is essential for enterprise use cases, such as industrial automation and mission-critical comms.

2. Predictive and autonomous operations

This operating process allows the automated detection and resolution of service degradation trends of all monitored digital services, and requires an innovative approach that involves:

• Leveraging machine learning based forecasting, anomaly detection, root cause analysis, and pattern detection, which are foundational for SLA guarantees and proactive service assurance.
• Strategic use of agentic AI leveraging best in class LLMs as the ultimate decision engine for the auto-remediation action and closing the automation loop.
• Use the Autonomous Networks (AN) framework and TM Forum’s 1 to 5 maturity model for a structured approach to automating operations and ultimately use generative AI as the business intent translation engine on the way to AN level 5.

3. Humanized intelligence with generative AI

The new operating model is not only about customer centricity and automation, but is also about speed, agility and simplicity. And this requires a new revolutionary way of accessing and interacting with data.

Generative AI capabilities can transform the way network intelligence is democratized across CSPs’ business and operations teams. By using natural language queries, CSP execs and staff as well as partners and customers can be empowered to “talk to the network and services” and get instantaneous insights enabling quick decision making.

A generative AI platform providing a natural language interface can support business operations by:

1. Cutting data availability from days to seconds.
2. Providing comprehensive insights to support strategic business decisions about network investments and new services by combining business, technical and customer data.
3. Enabling simplified and faster decision making.

An intelligence engine underpinning the new operation model

The three pillars of the transformation described above are dependent on the ability to maintain end-to-end visibility of customer, service and network experience in near-real-time and to automate at scale using AI. This cannot be achieved with traditional OSS/BSS approaches where data is organized in silos, inventory and topology information outdated and operations heavily manual.

CSPs need to build a new AI-native intelligence engine capable of generating these insights and commanding the automation actions. But AI efficacy, as we know, depends on data quality. This engine must come with a strong data foundation where high quality flattened, correlated and normalized data can ensure quality model training and model inference. Without this data foundation, an Intelligence Engine would struggle with noise, inconsistency, and false inferences.

So, the 3-phase approach to building the Autonomous Networks Intelligence Engine is:

• Build the data foundation by using multi-sourced data, both structured and unstructured.
• Utilize Agentic AI at its core and leverage the power of cloud and LLMs.
• Infuse the domain knowledge to turn the intelligence engine into the brain for close loop automation and Autonomous Networks.