Artificial intelligence to improve our services

“Who is not familiar now with Siri (Apple) or Alexa (Amazon) or has not seen Google Assistant on their mobile phone?”

In the case of Alexa, things go a step further because Amazon has launched Echo in Spain, the intelligent speaker that connects to its virtual assistant so that it can interact with her from any location in the household.

Today it is possible to ask Alexa using your own voice to play music, read a book or tell you tomorrow's weather. But furthermore, thanks to the new Endesa skill, it is also possible to ask about your energy and have tailored energy advice that enables saving on the bill closer than ever. And the only thing needed is being an Endesa client and having one of the different Amazon Echo wireless speaker models.

Thanks to this, at Endesa we offer a customer service channel by means of a device that enables the customer to control all the home automation. By combining both factors, we coordinate our customer service with information on consumption and bills together with the possibility of controlling the home’s temperature, lighting, etc. Therefore, this is an all-inclusive service that empowers the customer and gives them more decision capacity on saving and makes energy expenditure more efficient.

Our intention is to continue to make progress on this development so that by using their voice customers can carry out the same dealings or queries as they have been able to do up to now on the company's website or app. In future Alexa can even be used to open up communication with the contact centre in case additional service is required.

Moreover, we will continue to develop new voice channels to be able to reach all our customers. Along these lines, we are preparing a “skill” for Google Home which we hope will be up and running soon.

At Endesa we are also using Watson, IBM’s artificial intelligence to help customers both in the chat and by telephone and we do so with a significant percentage of clients attended to autonomously; all of this is to develop new solutions and optimise costs more quickly. After a pilot experience in 2017, last year it was extended to the entire customer service centre and currently helps customers both on the chat and telephone with a significant percentage of people attended to autonomously.

Intimately related to AI and the customer, at Endesa we are also working along these lines:

  • Models to prevent non-payments: creation of machine learning analytical solutions to reduce the number of non-payments and increase the charging rates in the non-payments portfolio.
  • RPAs. Data coder robots.
  • Complaints classifier. CRM complaint classification based on text recognition.
  • MEGABAT: Forecast of time demand for domestic customers that enables having stable consumption series over time, the most independent of commercial activity, customer ins and outs, thereby achieving savings and more speed in the service.

“The forecast of time demand, the reduction of the number of non-payments, the CRM complaint classification and the data coder robots are some of the objectives of the commercialization”

IA in distribution networks

We are also working to make progress on digitalisation of the distribution network by means of incorporation of artificial intelligence techniquesbig data in energy, and thereby improve our services.

Therefore, PASTORA was set up; the project to improve real-time control and preventive maintenance of the distribution network that carries electricity to homes and that will use big data, deep learning and artificial intelligence technologies to use the millions of data offered by the intelligent network. The aim is to develop predictive models of the network’s behaviour and therefore improve its operation.

The big data has reached the electrical network and will change everything. The millions of data offered by intelligent networks will enable system operators to predict and anticipate possible incidences to improve the network’s operation and increase the quality of customer service.

Within this digitalisation process, at Endesa we have also implemented the LARS system (Breakdowns and Re-establishment of Supply system), an automatic system that operates when a breakdown occurs as a virtual operator and performs from the same control centre system - just as human operators would do - operations on the network necessary to isolate breakdowns and re-establish supply in less than three minutes. LARS is responsible for remote management and will do so much more in the future.

Work is in progress to integrate historic data from breakdowns, failure rates, construction maps of streets and weather forecasts among other information so that the system can determine more accurately at what point in the line the breakdown occurred. Using big data and Artificial Intelligence techniques to process all this information, for example where lightning struck, where construction work on the street that may affect a cable is being carried out, the likelihood of determining the origin of the incidence more quickly increases significantly.

“La inteligencia artificial también ha cambiado la distribución de la energía: desde el aumento de velocidad en la localización de averías hasta la creación de algoritmos para reducir el fraude eléctrico.”

We also use artificial intelligence to respond to fraud. We are developing an algorithm that enables detecting fraud more reliably. This algorithm, responsible for Machine Learning, has been applied since 2017 all over Spain. The application of Machine Learning and Deep Learning algorithms on Big Data is enabling many of the company's operations to be approached more efficiently including the response to fraud.

IA en la generación

Y si la lA nos ayuda a atender mejor a nuestros clientes y a avanzar en la digitalización de nuestra red de distribución, también la estamos incorporando en proyectos de nuestra línea de negocio de generación.

  • Diagnóstico Predictivo: Empleo de técnicas de machine learning y análisis de datos operativos para la detección y diagnóstico de fallos en equipos principales de planta, con el fin de reducir la indisponibilidad y evitar averías.
  • Cámaras de supervisión con visión artificial: Desarrollo de modelos visuales inteligentes mediante el uso de redes neuronales para la identificación automática de eventos y problemas operativos.
  • MOP: proyecto piloto para la introducción de un sistema de ayuda a la operación de plantas de generación basado en inteligencia artificial y modelos cognitivos.
  • Monitorización de motores e interruptores de baja y media tensión: Empleo de sensórica avanzada y algoritmos de Inteligencia Artificial para la detección precoz de fallos en componentes eléctricos.
  • Digital Report: Desarrollo e implementación de un sistema global único para generación de informes, con acceso a los datos de las aplicaciones principales de generación e incorporando un asistente digital para información por voz.
  • AI4OM – Artificial Intelligence For O&M: Prueba de concepto para el desarrollo de una herramienta de mantenimiento prescriptivo, capaz de predecir el grado de fiabilidad de los eventos detectados por herramientas predictivas y sugerir en automático una actuación de mantenimiento adecuada a la situación.