Telecom Argentina speeds up technical incident resolution using Google Cloud tools

Telecom Argentina , a company providing connectivity experiences, entertainment, and tech solutions in Argentina, decided to start using Google Cloud as part of its hybrid multi-cloud strategy under the cloud-first concept.

With over 30 million customers, the company is moving ahead to fully digitalize its operations. That is why it has chosen to work with world-class partners that contribute with their knowledge and experience to its various transformation efforts.

Google Cloud was chosen for its robust data analytics and management and artificial intelligence offerings.

The first steps toward solution management

In early 2021, Telecom Argentina decided to start a disruptive project: creating an AI framework for its IT operations (AIOps) to increase the efficiency of its network assurance, QA, and customer service teams.

To this end, the company needed to streamline processes, replace on-premises models, and incorporate new technologies, focusing on a cloud-first strategy. Therefore, it decided to build a minimum viable product (MVP) for AIOps from scratch in Google Cloud, which would allow Telecom to improve its speed, agility, and time-to-market.

The AIOps project took about nine months to complete, during which aeros , a Google Cloud Partner in Latin America, played a crucial role in monitoring the solutions’ development.

The first technical challenge was incorporating 3 TB of data in just six months and gradually adding the same volume in datasets to create a data warehouse capable of managing an algorithmic machine-learning model with 329 estimators (trees) and 450 explanatory variables.

Currently, Telecom Argentina processes millions of data points after incorporating solutions such as BigQuery and Cloud Storage to store information; Vertex AI to build prototypes for ML models; and Cloud Functions to streamline service connections.

This has not only led to efficient data analytics, but also helped increase internal circulation to make processes more agile. Telecom Argentina now processes 3.8 TB of data every day for that project alone.

According to Righetti, this is not just a benefit in the form of more agile processes, but also "a way to democratize data" and make information more accessible to work teams.

Technical deployment challenges

In an early phase, a team of aeros, Google Cloud, and Telecom Argentina experts designed an architecture suited to the company’s scaling needs.

The team in charge of the AIOps project quickly began building an ML model to provide reliable assessment metrics for customer claims made in the contact center.

On the other hand, in the deployment phase, the team was able to detect certain difficulties in customer incident management that could lead to misunderstandings.

Google Cloud’s tools enabled the experts to address and quickly solve these data science issues comprehensively and iteratively.

Early results with Google Cloud

Although Telecom Argentina’s project is still in the development phase, the company obtained significant benefits using Google Cloud’s solutions.

Firstly, thanks to Google Cloud’s technologies and aeros’s consulting and monitoring, the company can now perform tasks that used to take a great deal of effort and several hours far more quickly.

The tools’ instant availability enables quick testing and immediate results with no need to request special permissions or access to specific platforms.

Moreover, the company is now capable of running multiple simulations concurrently, computing around 500 GB of data per case. To do that, it has virtual machines with high computing power providing results and insights on potential improvements to the model in just a few hours.

After deploying Google Cloud’s solutions, Telecom no longer needs to focus on maintaining or purchasing physical infrastructure, reducing both its capital and operational expenditures.

Consequently, the company is quickly and consistently improving its predictive model and can provide continuous business enhancements to focus not just on developing and evolving its model, but also helping internal customers with decision-making.

The new solutions have also allowed Telecom to shorten response times to internal and external customers. The technology team has streamlined its connection with customer service operations, and, in turn, they can speed up their response to the company’s end customers.

The project is expected to provide even more significant impacts on the business as it progresses, such as an improvement in Telecom’s net promoter score (NPS), a customer satisfaction indicator, thanks to the customer service enabled by the results from the models.

Thinking about the future

Currently, Telecom Argentina is working with aeros to improve data transfer and computing processes and start outlining a model with real-time operations.

The company is currently deploying in-house information processing capabilities using BigQuery, with an eye on future projects.

Telecom is also assessing additional Google Cloud solutions to optimize and automate existing processes so it can improve the quality of its customer service. As highlighted by Righetti, "Google Cloud is a strategic partner, not a provider."