Over the last few years, Metricell’s software has helped network operators handle millions of customer trouble tickets, arising from a variety of network-related issues. Despite the range of customer reported problems, be it call drops, slow data speeds or poor mobile signal, the root causes typically boil down to known incidents, maintenance activities, or underlying problems with coverage. Network operators face the formidable task of managing complex networks while addressing a constant stream of customer-reported issues.
For network teams, Artificial Intelligence (AI) offers new capabilities that allow them to spend less time reacting to issues and more time enhancing the network.
Today’s challenges
Manual Data Collection and Analysis
Currently, root-cause analysis involves collecting data from various sources, including live network status, maintenance tickets, incident reports, OSS performance stats and many more. This data is manually analysed by technical teams to identify patterns and pinpoint the source of network issues. The process is often fragmented, requiring significant time, access to multiple tools and the expertise to correlate and understand the data in depth.
Reactive Approach
Most network issues are handled reactively. When a problem is reported, engineers start investigating, which involves gathering data, analysing, and troubleshooting. This reactive approach often leads to longer resolution times and customer dissatisfaction due to delays and prolonged service disruptions. Although new technologies like AI are being introduced for predictive maintenance, many operators will still be reacting to subscriber complaints for years to come.
Limited Integration and Automation
Traditional root-cause analysis methods struggle with poor integration between different data sources. Network performance metrics and customer feedback are often kept separate, making it hard to get a complete view of the issue. Additionally, the lack of automation means much of the analysis is manual, which is time-consuming and resource intensive.
Where AI is assisting operators today
Proactive and Predictive Maintenance
AI enables a proactive approach to network management, by continuously monitoring network performance. It can predict potential issues before they impact customers allowing operators to address problems sooner, reducing downtime while enhancing customer satisfaction.
Enhanced Automation and Efficiency
AI is rapidly reducing the need for manual analysis and is quickly able to drive new automation, providing operators can integrate AI powered tools into their existing processes. AI is offering new ways to correlate data from diverse sources.
Tailored Insights for Stakeholders
Outputs from AI can be customised for different internal stakeholders. For example, Engineers may receive detailed technical reports to troubleshoot, while Customer Support Agents may get concise summaries that help them communicate effectively with customers.
Further automating network investigation steps with internal AI-powered systems
AI is only as informative as the data that goes into it. Any AI-powered solution starts with the data. In the context of RCA within telco, the starting point is to integrate data from multiple sources into a single, cohesive platform of data lake. Metricell’s Automated Network Investigator (“ANI” for short) is fed the data required to check and analyse:
- Live network status for real-time updates on network performance and health
- Incidents and maintenance on serving sites past, present and future
- The distance and line of sight to Cell Sites serving the area and what the predicted signal levels are by frequency band
- KPIs relating to network availability, accessibility, quality and others as required
- If other subscribers have reported issues in the same area
- Telephony measurements reported via crowdsourcing to show service availability, coverage levels and data speeds reported by subscriber handsets in the area
- Data from any AI-powered self-serve applications, such as Metricell’s AIDA
Results are reviewed and analysed automatically by AI to provide a range of concise and accurate descriptions of the potential problem root-cause. Several response types are generated for various stakeholders including:
- Scripted Response for Subscriber / Level 1 Support: Brief, concise acknowledgement of problem report with focused response pinpointing any obvious causes for the customer problem. This report typically automatically sent directly to customer phone, email or text.
- Assisted Response for Level 2 Support: Intermediate level information reporting on local network status issues, site performance testing and other tests which fail and relate to the customer problem report.
- Engineer Response for Level 3 Support / Engineering Teams: Tailored for technical teams, the Engineer Response delves into the specifics of the diagnostics performed, including what analysis was performed, detailed KPIs and technical findings. It will summarise the tests conducted across each network context area.
The traditional methods of root-cause analysis in telecommunications are becoming increasingly inadequate compared to the level of network complexity. AI is going to bridge the gap and tools like Metricell’s Automated Network Investigator, ANI, offers a transformative solution that automates and enhances the root-cause analysis process. Embracing AI will empower mobile operators to stay ahead of the curve, ensuring superior network performance and exceptional customer experiences over any competitors that don’t innovate.
Contact us here to discuss this topic in more detail and how it may apply to your operation.
