Apr 19, 2021

Playing to Win with AI for 5G Business Assurance

Try HeavyIQ Conversational Analytics on 400 million tweets

Download HEAVY.AI Free, a full-featured version available for use at no cost.



OmniSci’s no-code visualizations help Telcos optimize live 5G spectator experiences. Rapid iterations over massive fused geospatial datasets empower network providers to transform potential churn events into customer loyalty opportunities 

As 5G networks roll out globally, Communications Service Providers (CSPs) are positioning to deliver innovative customer experiences. One area CSPs are targeting for advanced 5G services is live sporting events. The expanded capacity promised by 5G has the potential to deliver astonishing experiences that incorporate real-time text, voice, high-resolution video, and not to mention augmented reality. 

Consumers are greeting these premium 5G services with enthusiasm and high expectations. Fans have come to expect seamless network connectivity, the ability to check scores, enjoy simultaneous live and on-screen action, all while connecting with friends to share highlights.  If a sports league decides to offer 4k or 8k multi-angle, in-stadium view, the CSP must deliver the expected customer experience.  In response to this demand, venue operators are looking to optimize network capacity and ensure that these new experiences live up to customer expectations.

Sporting venues present a unique challenge for 5G networks. On one hand, stadiums and arenas are used only infrequently. Operators lack the continuous data they typically use when planning networks that have predictable, day-to-day bandwidth usage patterns. On the other hand, no one can predict the trajectory of a game or even a season. As such, service providers need to be strategic in their 5G investments, adopting a different and more intelligent approach.

One way to do this is to embrace business assurance capabilities, which enable CSPs to guarantee that they are delivering the service levels that customers expect. 5G networks that embrace business assurance can enhance customer experience (CX), increase average revenue per user (ARPU), and optimize customer retention in the telecom industry.

To meet the new challenges of 5G, CSPs need to transform their business assurance solutions to support 5G’s inherent dynamism and complexity.  Along with AI and machine learning, CSP’s need to evolve their ability to quickly visualize huge datasets and simulate a wide variety of possible scenarios. 

5G networks generate huge quantities of data, which, to be effective, must be fused with other datasets, in particular, geospatial datasets. Geospatial data can be used to analyze spectator patterns at specific venues or events, providing insights into customer behavior. 

Big data analytics can certainly help with capacity planning and business assurance, but traditional BI visualization solutions have not kept pace with the rampant growth of data. At the same time, geospatial platforms are struggling to scale. Data analysts expect answers immediately, and they simply ask that BI tools “just work”.

Demonstrating Business Assurance at the TM Forum

In 2020, the TM Forum Virtual Leadership Summit offered the opportunity to demonstrate how powerful, real-time geospatial data visualization can empower CSPs to align spectator experience with business goals for 5G services. 

The TM Forum demo was built around data from a football match at Parken Stadium in Copenhagen, simulating an event attended by over 50,000 spectators. OmniSci’s Principal Solutions Consultant Eric Kontargyris presented the demo, sharing a number of custom, interactive dashboards that demonstrate the ability to rapidly visualize large, heterogeneous datasets. Working with a number of technology partners to provide a complete solution, OmniSci Immerse provided the visualization component of the demo. 

OmniSci’s real-time dashboard displays wireless customer and network data points for 50,000 spectators at a football match.

Visualizing the Dynamic Network Behind 5G Experiences

Collectively, 50,000 fans make heavy use of wireless services before, during, and after a match. Data usage fluctuates dramatically as fans connect with each other, check-in, and share pictures, tweets, and video over social media. Post-game, fans search for restaurants, plan routes home, replay game highlights, or connect with athletes’ premium post-game content. The rising usage of a second screen with high resolution live streaming, multi-angle instant replay and virtual reality places a particular strain on network bandwidth. 

Crucially, the CSP needs to dynamically align network performance with tiered SLAs. Tiered SLAs are based on calculated customer value and estimated probability of churn. Studies have shown that customer churn is closely correlated with network outages and latency. For this reason, CSPs are eager to deliver top performance to their most valuable customers.

OmniSci Immerse displays subscribers by customer score in real-time.

To segment customers by value, CSPs rely on data partners to score users by bad debt risk, fraud risk, and customer churn risk. In this case, OmniSci relies on a data partner to aggregate and track those inputs to assign a customer value class to each of the 50,000+ fans.

Customer value scores are cross-referenced against a Quality of Experience (QoE) score, derived from a range of factors including latency, packet loss, and download rates. Using these inputs, another data partner assigns a QoE score and monitors it across the network, aggregating QoE by customer value. Every user’s QoE should stay above the minimum acceptable threshold. The highest QoE scores are reserved for VIPs.

On 5G networks, ‘network slicing’ enables this delicate balancing act. 5G communications create multiple virtual networks over a shared infrastructure. Each virtual network is a small slice of the overall available network, whose bandwidth SLA is tailored to each service. Dynamic network slicing is a powerful capability that enables CSPs to preserve customer experience by intelligently shifting subscribers across network slices in response to traffic spikes or network latency. 

In the TM Forum demo scenario, individual subscribers are placed into the appropriate network slice as they enter Parken Stadium. Technology partners assign each subscriber a churn risk assessment, a bad-debt score, and a fraud risk assessment. By combining these assessments, the  AI components of the solution  dynamically segments individual subscribers into the appropriate 5G network slices. The dynamic QoE scores are constantly monitored by AI, enabling the network itself to make decisions that optimize customer satisfaction.

The combined capabilities of these technologies deliver network behavior that is adaptive and intelligent. Yet, providers still need to understand network behavior under a variety of conditions. Using OmniSci Immerse, datasets from a single event can be replayed repeatedly to gain an understanding of ways the network might be made more resilient under varied scenarios. 

OmniSci’s mission is to make analytics instant, powerful, and effortless for everyone. OmniSci replaces the hodge-podge of ad-hoc tools that are typically used to process and present large augmented geospatial datasets in a meaningful way. The more efficient the visualization technology, the more iterations are possible within a limited time frame. More iterations, involving deeper visualizations, enable operators to experiment with more ‘what-if’ scenarios, leveraging rich datasets from individual events to prepare for a wider variety of outcomes. 

Traditional BI visualization solutions struggle to provide such agility, especially with large geospatial datasets. It’s common for data analysts to spend a significant amount of time simply waiting for query results, suffering from the all-too-common ‘spinning wheels’ phenomenon. As the visualization component of the TM Forum solution, OmniSci Immerse demonstrated rapid time-to-insight, unlimited by the size and complexity of the fused datasets. OmniSci in turn enables data analysts to interrogate massive datasets more quickly and more comprehensively than the current generation of data visualization solutions. 

Goal: Business Assurance

Network conditions evolve as the event progresses. Data usage spikes whenever the home team scores a goal, as spectators share the highlight and watch high-resolution replays from various angles. When such events occur, the sheer volume of 5G data makes it impossible for human operators to rapidly visualize their effects on the network and respond in time.  By providing next-generation user experience for data and insights, OmniSci Immerse empowered the overall solution to transform potential churn into customer loyalty opportunities, while also providing business value, such as incentives for customers to upgrade to enhanced services. 

Try for yourself today, download OmniSci Free, a full-featured version available for use at no cost.

To learn more about how OmniSci can help CSPs harness the power of machine learning and business assurance, visit https://www.omnisci.com/industries/telecom

View full demo here:


HEAVY.AI (formerly OmniSci) is the pioneer in GPU-accelerated analytics, redefining speed and scale in big data querying and visualization. The HEAVY.AI platform is used to find insights in data beyond the limits of mainstream analytics tools. Originating from research at MIT, HEAVY.AI is a technology breakthrough, harnessing the massive parallel computing of GPUs for data analytics.