Integrated AI/ML tools enable automated, prioritizes early detection of performance anomalies and offer granular forecasting capabilities that are tuned to specific network conditions and data.
Stockholm, February 24th, 2026. Polystar will debut new AI-based enhancements for the globally deployed Kalix automated assurance solution at Mobile World Congress 2026. The enhancements offer integrated anomaly detection and forecasting tools, based on innovations in AI, and will be available for demonstration at the Fira next month.
AI-Enabled Anomaly Detection
Kalix Anomaly Detection empowers operators with automatic detection, isolation, and prioritization of service-impacting anomalies across complex networks and different customer segments. Leveraging advanced AI/ML algorithms to monitor multiple time series KPIs simultaneously, the new extension automatically identifies, in real-time, deviations from expected performance with statistical precision.
The innovation accelerates Root Cause Analysis and draws on an integrated knowledge base and LLM-powered summarization tools – delivering contextual recommendations for remedies, together with incident reports and troubleshooting workflows – enhancing service performance and operational efficiency.
Capacity Forecasting
The new forecasting capacity tool, meanwhile, provides out-of-the-box capabilities for estimating and managing network node capacity. Building on Polystar’s DataOps processing functions and based on state-of-the-art MLOps techniques, the tool combines all ingested data sources, as well as the existing probe-based data collection, and applies proven ML models to derive long-term capacity forecasts and traffic predictions.
The predictions cover a wide range of access and core nodes as well as network-wide traffic forecasts, supporting a wide range of use cases. In addition to forecasts, automated AI-assisted alerts are dynamically provided, highlighting focus areas, so that operators can adopt a data-driven approach to upgrades before thresholds are approached.
The ML algorithms ensure that accuracy is enhanced through time, while data scientists can adapt and tune models to suit specific business needs. Users can select their own ML models and unique time series predictions can be created.
“Securing value from AI and ML is imperative”, said Mohammad Shaheen, Head of AI Solutions at Polystar.
“That’s why we’ve focused our innovation efforts on key pain points for operators while enabling users to fine-tune our enhanced capabilities to the specific needs of their businesses. Forecasting must be aligned with the reality of the network in question, while issues and anomalies have different impacts according to the service and customer base. Polystar’s innovations benefit all operators while enabling customization for each network and service portfolio.”
The new extensions to Kalix will be demonstrated at Polystar’s stand at Mobile World Congress (Hall 5, Stand 5F35).
