AI based root cause analysis engine to proactively identify, analyze and resolve critical performance and operational issues in the network.
Network complexity is manifold due to multiple technologies and multiple NEPs.
There are lot of disparate tools providing network information with little to no interaction.
Expert teams are needed to create complex business rules for issue alerting and correlation that need constant updation.
Non scalable logics, manual intervention for debugging causing high lead times.
CEMtics's CEMplifi 'Aggregation Engine' driven Data Silo unification.
CEMtics RCA algorithm leverage for AI based issue classification and characterization.
DLNN based root cause analysis.
Impact assessment analysis and scoring.
80% of issues automatically triaged and analyzed by the RCA Engine.
60% of Network Issues classified with Causatives: Coverage, Capacity, Operations.
30% of Issues deduced due to Adjacency Issues.
Scalable solution deployed using Spark with high response times.
Subscriber centric traffic demand forecasting accounting for customer growth factors across marketing and sales.
In the current environment of viral apps and hype competition, historical network traffic is no guise for future traffic growth.
How do we better incorporate our marketing and sales strategy into our network planning / forecasting process?
Traffic simulation engine built “ground up” using customer location and usage patterns. Analysis of type of customer demand per location leveraging Cemtics Location algorithm.
Multifactor dynamic-forecast of subscriber traffic leveraging CEMtics Churn and upsell propensity algorithms in addition to incorporating business inputs like new rate plans / service bundles.
90% accuracy in data traffic forecast in the next 6-12 months compared to 75% accuracy of traditional forecasting methodology. Deferment of network spend to next financial year making capital available for other urgent projects.
A Machine Learning based model to help mobile operators Quantify, Audit, Predict and Optimize energy consumption cost at cell sites, based on inventory, network usage and crowdsourced data.
What are the factors causing huge consumption difference between different cell sites?
How can I predict the energy cost and audit the bill received from the electricity company / tower vendors?
Prediction of energy consumption on cell towers with network and external data from crowdsourced sources using advanced Machine Learning algorithms.
Audit of electricity charges from tower vendor resulting in $ savings.
Sites with inefficient hardware w.r.t. power consumption flagged for upgrade to newer models.
Network parameter changes recommendation in non-peak usage hours to reduce energy consumption.