HT - with the Power of AI from Reactive to Proactive and Predictive

“Being the largest Telco in the Country and serving more than 3 million clients across different channels (mobile, fix, broadband and TV), is something that comes with a big responsibility.

Our focus is always on two things: bringing more innovation and value to our Clients and providing our services with the best possible Customer Experience.

In today’s modern IT environments we literally have tons of systems, microservices, on-prem and cloud solutions. Those are very challenging to operate or even ‘just’ monitor. Together with CROZ, we invested a lot of effort and knowledge to implement the AIOps platform and use the capabilities of Machine Learning to significantly reduce downtimes and improve our Client’s satisfaction by moving from a traditional reactive model to a more proactive and predictive one. And we did all that in just 6 months!

It was a real privilege and joy for our team to work with CROZ experts who again proved their professionalism, focus on problem-solving and capability to deliver the end result.”

-Đani Glavinić, Director of IT Operations and E2E QA Department

To improve customer service and enable new products, HT is continuously improving its IT infrastructure and investing in its experts’ knowledge in various ways. Also, microservices architecture is now the standard and has helped HT accelerate application development, increase system performance and availability, but it has also amplified the need for centralized logging and monitoring system. The goal was to provide better visibility, improve the high quality of existing services to HT Clients and minimize downtimes of their systems.

Traditionally, monitoring such systems consisted of manually setting custom thresholds for each service and monitoring various dashboards to know if systems were functioning properly. With complexity increase, it was difficult to keep track of the numerous dashboards, manually adjust thresholds with each application upgrade, and investigate why one part of the system was not working properly.

To support HT to increase availability and stabilize continuity of its core and critical services, we proposed using machine learning models combined with large amounts of log data to detect anomalies and determine the root cause of incidents in real time.

HT’s staff proved their technical excellence and provided absolutely crucial support during the entire project lifecycle.

The pilot phase verified that the CROZ AIOPS platform could detect both major outages and, more importantly, anomalies that were not detected with traditional monitoring. Such anomalies often meant difficulties for end users that would never be detected due to the large number of requests that are being served every minute. Followed by larger project implementation, the AIOPS platform was tailored according to specific clients’ business processes to ensure accurate detection and root cause analysis. The platform was extended to other core services to fully improve the monitoring of the customer’s IT infrastructure. Since Kibana dashboards were already in use, the platform was integrated to provide a consolidated single view through an already familiar tool.

Implementation of the CROZ AIOPS platform significantly reduced downtimes as it enabled the customer operations team to immediately focus on the core issue, rather than spending a lot of time and effort to discover the root cause of the problem. Apart from reducing downtime it actually saved a lot of effort (time and expenses) in identifying and resolving issues.

HT is now using the power of AI to provide top-quality service to their Clients with minimized downtimes. Additionally, with the AI platform in place, they are ready to easily handle the future workload in an optimized and automated manner ensuring that the new services are reliable, available, and secure.