Lessons From Huawei's CEM Approach for CX Leaders

Most telecoms operators know their networks better than they know their customers. They can tell you the average packet-loss rate on a cell site in real time; they struggle to tell you whether the person using that cell site is about to churn. Huawei's SmartCare Customer Experience Management (CEM) solution was built on the uncomfortable recognition that this imbalance is not a data problem — it is a strategic one. The lessons it surfaces apply well beyond telecoms.

The central argument here: effective CX management is not about collecting more data. It is about reorienting the entire measurement architecture — and the organisational logic behind it — around the customer's perception, not the operator's infrastructure. Huawei's SmartCare offers one of the most structurally rigorous examples of what that reorientation looks like in practice.

Why Network KPIs Are the Wrong Unit of Analysis

A telecommunications operator running on aggregate Key Performance Indicators (KPIs) is, in effect, managing a fiction. The network may be performing within acceptable thresholds while a specific customer in a specific location is experiencing a video call that keeps dropping. The aggregate looks fine. The customer is furious. This is the measurement gap that sits at the heart of most CX failures — not just in telecoms, but across industries.

Huawei's SmartCare addresses this directly through what it calls Per-Service Per-User (PSPU) monitoring: tracking service quality, identifying network issues, and determining root causes on a per-service, per-user, and per-raster basis rather than relying on network-wide averages. The implication for CX practitioners is significant. Aggregate metrics — whether network KPIs in telecoms or average CSAT scores in retail — systematically obscure the experiences of the customers most likely to leave. The mean is a comfortable number. The outlier is the signal.

This is behavioural economics in structural form. Daniel Kahneman's peak-end rule tells us that people judge an experience not by its average quality but by how it felt at its most intense moment and at its close. An operator whose aggregate KPIs are green but whose worst-served customers are experiencing repeated service failures is, in Kahneman's terms, delivering a peak that is negative and an end that is frustrating. Those customers will not remember the 95% of the time the service worked. They will remember the call that dropped during the job interview.

The Three-Tier Indicator Architecture: A Model Worth Borrowing

SmartCare's measurement framework is built on three distinct layers: Customer Experience Index (CEI), Key Quality Indicators (KQI), and Key Performance Indicators (KPI). The sequence is deliberate and instructive.

• CEI (Customer Experience Index) sits at the top — it is the customer's perceived quality of service, the number that matters to retention and loyalty decisions.
• KQI (Key Quality Indicators) sit in the middle — service-level metrics such as call setup success rate or video streaming resolution that translate network behaviour into service terms.
• KPI (Key Performance Indicators) sit at the base — the raw technical measurements of network infrastructure performance.

The architecture forces a diagnostic logic that runs from customer perception downward, not from network performance upward. When a CEI score deteriorates, the framework prompts investigation at the KQI level, then at the KPI level. The customer experience is the starting point; the technical root cause is what you find at the end of the investigation.

Most organisations do the reverse. They build dashboards around what is easiest to measure — operational metrics — and then attempt to infer customer experience from those numbers. The inference is almost always wrong, because operational efficiency and customer perception are related but not equivalent. A bank can process a complaint in four hours (a strong KPI) while making the customer feel unheard throughout (a collapsed CEI). Designing a Voice of Customer strategy that mirrors this three-tier logic — starting with perception, working backwards to process, then to system — is one of the highest-leverage structural changes a CX team can make.

Journey Mapping as Operational Infrastructure, Not Workshop Output

SmartCare maps customer journeys across four touchpoint categories: Interact, Choose, Use, and Get Help. These are not conceptual labels for a workshop slide. They are the operational categories around which cross-departmental collaboration — between marketing, network planning, and customer service — is organised.

This is the distinction that separates CX management programmes that produce change from those that produce documentation. Journey maps are valuable only when they are connected to the systems and teams that own each stage. A map that lives in a PowerPoint deck and is reviewed quarterly has no operational consequence. A map that is embedded in the measurement architecture — where each stage has defined indicators, owners, and escalation paths — is a management tool.

The Get Help category deserves particular attention. It is the stage most operators and service businesses underinvest in, and the one with the highest emotional stakes. Research by Dixon, Freeman, and Toman, published in Harvard Business Review in July 2010, established that reducing customer effort in service recovery is more predictive of loyalty than delighting customers at the primary service stage. Customers who experience a problem and have it resolved effortlessly are often more loyal than customers who never had a problem at all. The Get Help touchpoint is not a cost centre to be minimised; it is a loyalty lever to be engineered.

For organisations building or rebuilding their customer journey architecture, the SmartCare framework offers a useful structural prompt: have you mapped the Get Help stage with the same rigour as the primary service journey? And does that map connect to real operational owners?

Multi-Vendor Agnosticism as a CX Principle

SmartCare is vendor-agnostic: it collects data directly from interfaces via probes to integrate infrastructure from multiple vendors — Nokia, Ericsson, ZTE — onto a single platform. This is a technical capability, but it encodes a CX principle that is widely violated in non-telecoms contexts.

The principle: the customer does not care which vendor, system, or department owns a given touchpoint. Their experience is continuous. The organisation's internal boundaries are invisible to them and irrelevant to their perception. When a customer's complaint falls into the gap between a CRM system and a billing platform because the two do not share data, the customer experiences a single failure — even if the root cause is a technical integration problem between two vendors' systems.

The practical implication is that CX management must be architected at the level of the customer's journey, not at the level of internal system ownership. This requires both technical integration (the SmartCare approach) and organisational integration — governance structures that assign accountability for the full journey rather than for individual departmental touchpoints. Without the latter, even the best measurement architecture produces data that nobody has the authority to act on.

What the Customer Experience Transformation Centre Tells Us About Organisational Readiness

In 2014, Huawei established the Customer Experience Transformation Centre (CETC) — a dedicated environment designed to build a collaborative CEM ecosystem and verify business value for operators. The existence of this structure is as instructive as the technology it supports.

Huawei recognised that selling a CEM platform to a telecoms operator was not sufficient. The operator also needed to develop the organisational capability to use it — to interpret CEI data, to act on KQI signals, to coordinate across departments that had historically operated in silos. The CETC is, in effect, a change management mechanism embedded in the product delivery model.

This is a lesson that applies directly to any organisation implementing a CX management programme. Technology is rarely the binding constraint. The binding constraints are almost always organisational: unclear accountability, siloed data ownership, a culture that rewards operational metrics over customer outcomes, and leadership that has not yet internalised the shift from efficiency management to experience management. Change management is not a phase that follows CX implementation; it is a prerequisite for it.

The IKEA effect — the behavioural tendency to place disproportionate value on things we have helped to create — is relevant here. Organisations that co-design their CX measurement frameworks, rather than receiving them as a vendor-configured package, develop stronger ownership of the outputs and greater willingness to act on uncomfortable findings. The CETC model, which involves operators in building the ecosystem rather than simply consuming it, appears to encode this principle deliberately.

AI-Native CX Management: What the 2026 SmartCare Intelligence Launch Signals

At MWC Barcelona 2026, Huawei launched SmartCare Intelligence — an evolution of the platform introducing an AI-Native architecture designed to accelerate autonomous network operations toward Level 4 autonomy. Two features are particularly relevant to CX practitioners thinking beyond telecoms.

The first is DataChat: a unified telecom-domain AI entry that uses long-term memory and domain knowledge to proactively analyse network traffic and user experience. The key word is proactively. Most CX analytics systems are retrospective — they tell you what happened after the customer has already formed a judgement. A system with long-term memory that surfaces emerging experience degradation before the customer notices it represents a structural shift from reactive to predictive CX management.

The second is Experience Digital Twins (EDT): the integration of network configurations with real-time experience data to create digital twins of network and customer behaviour, enabling millisecond-level monitoring for 5G and 5G-Advanced services. The digital twin concept — a live, dynamic model of the customer's experience that runs in parallel with the actual experience — is one of the most significant architectural ideas in CX management. It allows organisations to simulate the impact of operational changes on customer experience before those changes are deployed, rather than discovering the impact through post-hoc complaint analysis.

For organisations outside telecoms, the practical translation is this: the direction of travel in serious CX management is toward real-time, predictive, and simulation-capable systems. Organisations still operating on quarterly survey cycles and annual journey-map reviews are not just behind — they are measuring a different game.

Five Transferable Lessons for CX Leaders

The SmartCare framework is telecoms-specific in its implementation but universal in its logic. The following lessons transfer directly to any organisation serious about CX management (CX management) as a discipline rather than a function.

1. Reorder your measurement hierarchy. Start with customer perception (your equivalent of CEI), work backwards to service-level indicators (KQI), then to operational metrics (KPI). Never let the operational layer drive the customer experience agenda.
2. Disaggregate your data. Averages protect underperformance. PSPU-style monitoring — tracking experience at the individual customer level, not the segment or cohort level — surfaces the signals that aggregate reporting buries. The customers most at risk of churning are rarely visible in the mean.
3. Connect journey maps to operational owners. A journey map without an accountability structure is a design artefact. Each stage of the journey — particularly Get Help — needs a named owner, defined indicators, and a clear escalation path.
4. Architect for integration, not for departmental convenience. The customer's experience is continuous. Your internal systems and governance structures need to reflect that continuity, not fragment it along departmental lines.
5. Treat organisational capability as a co-deliverable. Technology enables CX management; it does not deliver it. The human and organisational capability to interpret data, make cross-functional decisions, and act on uncomfortable findings must be built in parallel with any platform implementation.

The Deeper Strategic Lesson: From Network-Centric to Human-Centric Operations

Huawei describes SmartCare's purpose as enabling telecoms operators to transition from network-centric operations to service- and user-centric operations. That sentence is worth sitting with. It is not describing a technology upgrade. It is describing a strategic reorientation — a change in what the organisation believes its job is.

A network-centric operator believes its job is to run a reliable network. A user-centric operator believes its job is to deliver reliable experiences to specific people. The network is the means; the experience is the product. This distinction sounds obvious when stated plainly. It is, in practice, resisted by almost every organisation that has built its identity, its metrics, and its incentive structures around operational excellence rather than customer outcomes.

The same reorientation is available — and necessary — in every service industry. A bank that believes its job is to process transactions efficiently will build different systems, hire different people, and measure different things than a bank that believes its job is to give customers confidence in their financial decisions. A hospital that measures bed-occupancy rates as its primary performance indicator will make different decisions than one that measures patient experience across the full care journey. The CX maturity of an organisation is, at its core, a measure of how completely it has made this reorientation — not how sophisticated its technology is.

SmartCare's three-tier indicator system, its PSPU monitoring, its journey-based operational framework, and its AI-native evolution are all expressions of a single strategic commitment: the customer's perception is the primary unit of management. Everything else — the network, the systems, the departments — is infrastructure in service of that unit.

That commitment, once genuinely made, changes everything downstream. The measurement architecture changes. The governance structure changes. The conversation in the boardroom changes. The technology choices follow naturally from the strategic orientation, not the other way around.

Organisations that are still trying to infer customer experience from operational metrics have not yet made the commitment. They are, in Huawei's framing, still network-centric — whatever their CX strategy document says. The question for any CX leader is not whether their organisation has the right platform. It is whether the organisation has genuinely decided that the customer's perception is the primary unit of management — and whether every measurement, governance, and accountability structure reflects that decision.

If the answer is not yet, the technology is the least of the problems. For organisations ready to close that gap, exploring a structured customer experience management engagement is a practical starting point — not to acquire a framework, but to pressure-test whether the strategic commitment is real.

Frequently Asked Questions

What is customer experience management (CX management)?

CX management is the discipline of systematically designing, measuring, and improving every interaction a customer has with an organisation — across all touchpoints and throughout the full customer lifecycle. It differs from customer service in scope: where customer service addresses individual interactions, CX management governs the cumulative experience and the organisational structures that produce it.

How does Huawei SmartCare CEM work?

Huawei SmartCare CEM collects data directly from network interfaces via probes, integrating infrastructure from multiple vendors onto a single platform. It uses a three-tier indicator system — CEI (Customer Experience Index), KQI (Key Quality Indicator), and KPI (Key Performance Indicator) — to translate raw network metrics upward into customer-perceptible quality scores. The CEI sits at the top of that hierarchy: it is the number the organisation manages to, and the KQIs and KPIs beneath it explain why it is moving. That architecture is precisely what makes the system customer-centric rather than network-centric. The platform also enables segmentation by subscriber, service type, and location, so degradation can be identified at the level of the individual experience rather than the aggregate network average.

What can CX leaders outside telecoms take from Huawei's approach?

The transferable principle is the indicator hierarchy itself. Any organisation can audit whether its measurement stack is built top-down from customer perception or bottom-up from operational convenience. If the primary metrics are internal — handle time, resolution rate, system uptime — and customer perception is inferred from them, the architecture is inverted. Correcting that inversion does not require a new platform; it requires a deliberate decision about what the organisation is actually trying to manage, followed by the discipline to build every measurement, reporting, and accountability structure around that answer.

For CX leaders in sectors beyond telecoms, the honest starting point is to map which metrics currently reach the boardroom and ask whether any of them directly represent how customers perceive value. If none do, the organisation has not yet made the strategic commitment that Huawei's CEM architecture embodies — regardless of what its customer experience strategy document says.