Internet Service Providers across the world are facing the common problem of trying to deliver the best customer experience and service assurance in the most cost-effective manner. Despite best efforts, the current tools and technologies deployed are not meeting those goals, and as a result service providers are struggling with low ARPU, high churn, and increasing support costs.
Ayla launches Analytics for Internet Service Providers
Ayla IoT has released a powerful new analytics-based solution, TransformAI, that applies the power of machine learning to big data gathered from devices in the customer home network. This new solution is unique because it replaces the traditional network-centric approach with a modern ‘customer-centric’ approach. The solution is focused on diagnosing, explaining, and predicting key service issues in the home network that are major contributors to adverse experiences and high support costs. By taking this 80/20 approach, the solution allows customer operations teams in leading service provider organizations to save upwards of 8% on annual customer support costs in year 1 alone, increasing to 12-15% in year 3.
Here’s a specific case study in which Ayla’s TransformAI solution was able to effectively address pressing customer experience challenges for a tier1 CSP with over 3.4M devices under management. The provider, which has millions of subscribers, was faced with the problem of high call rates and accompanying truck rolls with expensive device swaps that in many cases didn’t address the actual customer problem, leading to issue recurrence. Using Ayla’s solution, high volume data from millions of gateways, routers, and other CPE was collected, processed, and stored in a purpose-built data model. From there, the data was run against statistical models (for anomaly detection) as a first step to identify the issue and then fed into a model to attempt the predict the occurrence and root cause. Some of the problem categories that were analyzed included slow browse, router resets, sync-no-surf, and Wi-Fi authentication errors. To date, the models (developed, trained, and tested by Ayla data scientists) have drawn inferences with a high level of accuracy of over 85% allowing CX teams at CSPs to act early and avoid a scenario involving an irate customer call followed by needless truck rolls. Additionally, the solution is able to provide insights to trace problems down to manufacturer and firmware versions with quick slide-and-dice capabilities.
Direct, High-value Benefits to Stakeholders
The success of the Ayla TransformAI solution for communication service providers is now being replicated at other Tier1 and Tier2 entities. The key drivers of adoption are:
Rapid time-to-value – an average period of under 3 months to build the data management pipeline, analyze the data, and generate powerful insights even in a limited scope engagement
Agnostic and future-proof – because the solution is data-centric, it is not tethered to a specific class of CPE, it works across any device and network type within the home network, making it future-proof for any new devices added over time
High impact on the business – leveraging analytics has a direct, measurable impact on the business in terms of increase in ARPU and margin expansion
In a relatively short time period, various business functions including customer experience, customer service operations, network performance management, and innovation are seeing tangible benefits from harnessing data and deploying the customer experience-centric model that is contributing real savings. For the tier1 CSP referred to earlier, the OPEX savings amounted to $36M in the first year alone and are projected to increase to over $50M each in years 2 & 3.
Making predictive analytics work in a CSP environment requires deep domain expertise that comes from an understanding of how home network devices work at a deep technical level. That intangible aspect goes into training superior supervised models that achieve higher levels of accuracy and effectiveness. At Ayla IoT we possess over 80 person-years of combined Telco/CSP expertise that gives a distinct competitive edge when it comes to identifying issues, determining the root cause, and prescribing fixes. Ayla is actively rolling out new quantitative models for various CX problem areas through a repeatable, consistent machine learning operations (MLOps) system that is estimated to produce over 50 predictive models by the year-end.
Ayla TransformAI for CSPs is now generally available for tier1 & tier2 organizations with at least 500K subscribers. Contact us for a product demo or consultation and give us an opportunity to elevate your business performance.