Roughly 80% of operators already use AI for analytics; however, these deployments are largely limited to decision-support use cases
In partnership with Radcom, GSMA Intelligence has launched a new report, Service Assurance Trends in the AI Era, which revealed key focus areas for telcos in 2026 and documents a shift towards a more customer-centric approach through the use of unified assurance solutions, larger-scale agentic AI deployments, and real-time data infrastructure.
Top takeaways from the report:
End-to-end visibility remains a major gap despite “unified” platforms
While 82% of operators claim RAN-to-core correlation, only 41% have a true end-to-end data architecture across domains and departments. This disconnect limits real-time insight and automation potential, while 21% say they have a mix of both integrated architecture and siloed data management. The remaining 38% still have different teams or departments conduct data collection and management across their operations.
For GSMA Intelligence, this misalignment suggests that the majority of telcos may believe that they have visibility, but most still lack the data foundation required for AI-driven operations.
Service assurance is becoming customer-experience infrastructure, not just a network tool
Operators’ primary use of service assurance today is customer insight, churn reduction, and Net Promoter Score (NPS) improvement, not pure network troubleshooting. “Operators depend on assurance solutions to give them greater insight into customer usage and spending patterns, with a view to shoring up the overall customer experience and driving NPS,” stated the report.
Insights into customer and usage trends for upsell and marketing purposes (53%) and analytics on customer experiences to improve experience and reduce churn (52%) make up the top two use cases for network data, followed by network use cases: analyzing network problems on the core (48%) and for advanced network troubleshooting (42%).
The takeaway is that more than half of operators use network data primarily for customer analytics and monetization, highlighting how assurance has shifted from OSS plumbing to a frontline customer experience system.
Here, GSMA Intelligence also pointed out another “disconnect,” this time between assurance and OSS/BSS. “The survey points to a disconnect in that the multiple streams of data available to operators from analysis of their customers and parts of the network are not necessarily integrated with and streamed to existing OSS/BSS,” it said. “While this may be work in progress, it could suggest a gap or opportunity for existing assurance platforms to fully integrate with these systems.”
AI is already mainstream — but mostly assistive, not autonomous
About 80% of operators already use AI for analytics such as anomaly detection, fault prediction, and root-cause analysis. However, these deployments are largely limited to decision-support use cases, including customer complaint analysis (63%), fault prediction (59%), and root-cause analysis (55%), rather than closed-loop, autonomous automation, which just 39% have implemented.
These findings show that while telecom is indeed injecting AI into its operations, it has yet to do so in an AI-native manner.
Agentic AI is the next inflection point
Roughly 65% of operators see agentic AI as transformational or high-value, and over 70% of those planning deployments target 2026. Priority use cases include automated complaint resolution (57%), autonomous fault resolution (54%), and customer experience predictions (52%).
Additionally, nearly half of the respondents (48%) said that agentic AI would be valuable for predictive network health scoring.
Mobile data quality is now the dominant pain point
45% of service complaints relate to mobile data, especially streaming and gaming, while voice generates minimal complaints. The low level of complaints relating to voice is not surprising,” noted GSMA Intelligence, explaining that voice is a “stable product segment” and “highly optimized” on the network. “Data is not, however,” the report continued. “Depending on the specific operator, country, coverage areas, and so on, data coverage and capacity can vary significantly, generating the highest level of complaints.”
However, most operators (71%) claim to be handling these complaints well using existing systems. Even so, future assurance strategies will need to align more closely with data-centric services — including 5G standalone, network slicing, and emerging AI-driven workloads — to meet rising expectations for low latency, consistent throughput, and real-time performance guarantees.
Table of Contents
