Boosting NPS beyond network KPIs with agentic AI Your network is performing. So why is NPS still falling? The Challenge: The gap between a good network and a happy customer. Every CSP has invested heavily in measuring network quality. Customer Experience Indices built from hundreds of KPIs — coverage, throughput, latency, reliability — provide high-frequency, universal insight into how the network is performing. But boards don't track CEI, they track Net Promoter Score, and the link between the two is weaker than most operators expect. A customer sitting on a perfectly performing cell can still score you zero if their activation took three days, their billing dispute went unresolved for weeks, or a customer care representative put them on hold for forty minutes. Network quality is necessary, but it is definitely not sufficient. The Transformation: From reactive surveys to real-time intelligence This Catalyst moves operators from a world of lagging indicators and siloed responses to continuous, proactive, cross-domain NPS management. Before - Reactive NPS surveys find out a customer is a detractor weeks after the fact, when it's already too late to act. . Network-only models: CEI captures signal and speed but misses care contacts, billing disputes, and onboarding failures entirely - Siloed domains: Network, care, billing, and digital teams optimise independently, no unified view of customer experience - Manual response: Human-led diagnosis and intervention. Slow, inconsistent, and expensive at scale After - Continuous NPS estimation: NPS modelled in near-real-time from live network, care, billing, and digital signals without survey lag - Multi-signal intelligence: CEI augmented with every touchpoint that shapes customer perception; a complete picture of experience - Cross-domain correlation: Network, care, onboarding, and lifecycle data fused into a single NPS driver model per customer archetype - Agentic closed-loop action: AI agents detect NPS risk, identify root cause, and trigger personalised interventions autonomously Use Cases Each use case targets a distinct part of the customer journey where NPS is made — or lost. Together they form a closed loop from first impression to long-term loyalty. 1.Intent-driven service onboarding The first hours of a new service define the relationship. Agentic AI monitors the activation journey in real time detecting friction, adjusting parameters, and intervening proactively before a customer ever picks up the phone. Intent-based SLA management ensures the service delivered matches the service promised. 5G activation Intent management Onboarding NPS. 2. Proactive & personalised customer care Most care is reactive. This use case flips the model. By correlating network degradation, billing anomalies, and care contact patterns with NPS profiles by customer archetype, the system identifies at-risk customers and triggers personalised interventions before they complain, and before they churn. Churn prevention Proactive care NPS archetypes 3. Service lifecycle optimisation via digital twins Network planning decisions have NPS consequences — but those consequences are rarely visible when they're made. This use case integrates NPS estimation into a Digital Twin, so every planning decision (site deployment, sector configuration, upgrade prioritisation) shows its predicted customer impact. Architecture Overview: The solution is built on a modular, standards-aligned architecture: Data Layer * Network KPIs (coverage, latency, throughput, availability) * Customer care data (tickets, interactions, resolution metrics) * Incident and fault management * Service lifecycle and onboarding data * Device and service usage signals Intelligence Layer * AI/ML models for NPS estimation * Driver analysis and customer segmentation * Agentic AI for reasoning and decision-making Action Layer * Closed-loop automation across: * Customer care * Service onboarding * Network and service optimization This architecture enables continuous correlation, insight generation, and automated action across domains. TM Forum Alignment: This Catalyst is fully aligned with TM Forum frameworks and best practices: * IG1394 Telco NPS Management Framework v1.1.0 * Supporting Product, Service, and Network NPS correlation and optimization * Autonomous Networks Framework (AN Levels) * Enabling closed-loop, cross-domain automation * Open Digital Architecture (ODA) * Modular, interoperable system design * TM Forum Open APIs * Supporting integration across domains The project contributes to advancing AI-driven, customer-centric autonomous operations. --------------------------------------------------------------------- Business Impact This Catalyst enables CSPs to: * Transform NPS into a continuous, operational KPI * Identify and act on the true drivers of customer satisfaction * Break silos between network, service, and customer care domains * Proactively improve customer experience at scale * Reduce churn and increase customer loyalty Expected Impact: * +5–10% NPS improvement * 1–2% churn reduction * 20–30% improvement in customer care efficiency * >90% faster diagnostics and response times * Progress toward Autonomous Networks Level 3–4 For additional details, see featured resources below.