Energy and Utilities

Intelligent Automation to Optimize Customer Care

The Challenge

Our client, a reference company in the energy industry in Spain, provides essential services to thousands of customers, which involves managing a massive volume of data that is constantly growing. This is where the mission of its Customer Care Service takes on its full value, as its professionals process more than 50,000 incidents per year.

In most cases, these are service requests, complaints and follow-up, which are analyzed centrally by the team for classification and subsequent resolution. There is an added difficulty of having to manage false incidents, incidents with missing information and even many that correspond to other business units.

In this context, a significant part of the time was spent on manual classification of these incidents: a necessary job, although monotonous and subject to failures, and of little interest to the person performing it.

| The Outcome

At LAUDE we proposed the automation of this process with a double challenge: to speed up the incident resolution time and to make Customer Care professionals to dedicate their time to higher value-added tasks to optimize the user experience.

How we did it

We started our project with an important research effort: our consultants and data scientists designed a solution that was simple to implement, scalable and perfectly measurable in terms of KPIs and ROI.

After several prototypes, we chose to design a cloud-hosted system capable of automating incident routing, by means of two main tools:

A Natural Language Processing (NLP) model that analyzes the text of incident tickets and completes the output data to integrate it into a database.
Several Machine Learning (ML) classification algorithms that work against that database and assign the corresponding task to each type of incident: by automation possibilities, by typology, by department, etc.

However, ML and AI-based solutions were only the beginning: the most important thing was to establish a collaboration strategy with our client that would facilitate, in the long term, the improvement of the solution as the project evolves. To this end, an Analytics tool was developed to measure the results of the project and help us anticipate needs and future developments.