How Telefónica uses artificial intelligence and machine learning to connect the unconnected


Telefónica uses data science to systematically identify and locate the non-connected, including them in their networks and operations in order to connect most of Latin America in a sustainable way.


Rural connectivity Internet para todos


The “Internet para todos” project is the flagship of Telefónica to connect the unconnected in Latin America. Today, there are more than 100 million people living without a secure connection to the Internet where Telefónica operates. The reasons are multiple, ranging from geography, population density and socio-economic conditions.

Historically, fixed and mobile networks have been designed to achieve maximum efficiency in dense urban environments. The implementation of these technologies in rural and remote areas and low density is possible but inefficient, which challenges the financial sustainability of the model.

To provide Internet in these environments in a sustainable way, it is necessary to increase efficiency by means of the systematic reduction of costs, the optimisation of investment and specific implementations.

Systematic optimisation requires the continuous measurement of financial, operational, technological and organisational datasets.



1. Find the disconnected

The first challenge that the team faced was to understand how many people are not connected and where. The set of data was scarce and incomplete, the census was old and the population had a lot of mobility. In this case, the team used high-definition satellite images to the scale of the country, in addition to neural network models, along with census data and skills.

Implementing visual machine learning algorithms, the model literally counted each house and settlement in the scale of the country. Then, it was enriched with cross-referenced data from the regulatory source, as well as with the data set owned by Telefónica, which consisted of geolocated data sessions and implementation maps.

The result is a model with a visual representation, which provides a map of the dispersion of the population, with overlapping coverage, allowing us to count and locate disconnected populations with good precision (95% of the population with less of the 3% of false positives and less than 240 metres of deviation in the location of the antennas).


2. Optimise transport

The transport networks are the most expensive part of the implementation of connectivity in remote areas. The optimisation of the transport route has a great impact on the sustainability of a network. This is the reason why the team selected this task as the next challenge to face.

The team began by adding roads and infrastructure data to the model of public sources and used the generation of graphs to group together the population settlements. The analysis of graphs (shortest path, Steiner tree) provided optimised transport routes for the population density.


3. AI to optimise network operations

Connect very remote areas, streamline operations and minimise maintenance and updating are key to a sustainable operating model. This line of work is probably the most ambitious for the team. Because, when it can take 3 hours by plane and 4 days by boat to reach some locations, being able to be sure that it can be detected, or rather, predict if you need it or when you need to perform maintenance on your infrastructure is key.

Equally important is how to design their routes to make it as efficient as possible. In this case, the team built a trained neural network with historical failure analysis and fed with network metrics to provide a model capable of monitoring the status of the network in an automated way, with prediction of possible failures and an optimised maintenance route.

I think that the type of approach based on data for the resolution of complex problems demonstrated in this project is the key to the sustainability of the network operators in the future. It is not only a rural problem and it is necessary to increase efficiency and optimise deployment and operations to continue reducing costs.


Connectivity Latin America



Originally published in: