Published on13/10/2023

Unifiber x Deepomatic: Artificial Intelligence empowering fiber deployment in Wallonia

A first for the telecom sector in Belgium

Waterloo, October 13, 2023 – Deepomatic, the European leader in artificial intelligence applied to major infrastructure projects, and Unifiber, the company deploying the optical fiber network in Wallonia, have signed a collaboration agreement. The solution developed by both companies, leveraging artificial intelligence, enables automatic quality control of the work carried out by agents deploying optical fiber. Following a 5-month pilot phase, it is now ready for field use. This project marks a pioneering achievement in Belgian telecommunications.

Unifiber's projected numbers are staggering: by 2028, 600,000 households will be connectable to the new optical fiber network, a game-changer for those needing ultra-fast and stable connectivity for work or leisure. This entails installing approximately 10,000 kilometers of pink cables, on facades, along electrical pylons, and often underground. Quality control is a pivotal lever in Unifiber's strategy. The work is executed by numerous teams, each with their own quality requirements. To align all stakeholders to a consistently high quality standard, Unifiber has partnered with Deepomatic.

Deepomatic, a French company founded in 2014, has since seen its solution adopted by major telecom and industry players worldwide for managing the deployment of various infrastructures: Bouygues Telecom in France, Swisscom in Switzerland, Movistar in Colombia, and Circet in Ireland.

Real-time Quality Control

Deepomatic's computer vision solution streamlines on-field team quality control. It analyzes images captured by the teams in real time, providing immediate feedback based on result color codes. Green for a successfully completed task, red for a "non-compliant" checkpoint. This allows teams to swiftly address issues before leaving the site. Once the non-compliance is rectified, the team takes another photo that the system analyzes to demonstrate that the task was performed correctly. This procedure objectifies and systematizes large-scale quality control. The photos and analysis results are centralized and made available to the overseeing teams managing the deployment.

« By obtaining feedback on the quality of their work, teams can gradually improve their techniques. Moreover, the reduced error rate eliminates the need to plan additional visits. This efficient installation process reduces the risk of extra work and, consequently, delays. All of this contributes to a smooth and effective fiber deployment. When you deploy fiber, it's worth doing it correctly the first time. In other words: First Time Right. Deepomatic ensures the success of field operations from the first intervention. »

Philippe Jaspart, Quality Manager at Unifiber

For municipalities, this translates to operational excellence, increased success rates, fewer site revisitations, and higher satisfaction with the work carried out on the sites.

« At Deepomatic, we understand the importance of deploying sustainable infrastructure and optimizing field truck rolls to achieve operational excellence. Thanks to this project, Unifiber stands out as a pioneering player in the field of AI applied to optical fiber operations. »

Augustin Marty, CEO de Deepomatic

What Does Artificial Intelligence Control?

Deepomatic's mobile web application has been tested on sites since May with all of Unifiber's production partners. To enhance the application's performance, in other words, to make it more "intelligent," Unifiber has collected a sufficient quantity of photos. Today, the database contains over 5,000 "OK" and "KO" photos taken on sites during this period.

Deepomatic employs advanced computer vision algorithms to analyze the images captured on-site. These algorithms are trained to identify specific checkpoints and flag any inconsistencies or issues.

Currently, eight points are being controlled by artificial intelligence on Unifiber's sites. They validate the compliance of the trenches where the fiber is installed, the materials placed within (sand, fiber, signaling tape, backfill earth, coating and coating joints), the equipment constituting the infrastructure (such as fiber manholes), and the evenness of the ground.

New checkpoints are under consideration, including those for the installation phases of POPs (points of presence) and for household connections.

About Unifiber

Unifiber is a joint venture between Proximus and Eurofiber. Established in July 2021, it is based in Waterloo and employs around thirty people. The company can rely on two reference shareholders: Proximus and Eurofiber. Eurofiber is a major player in Europe for the development of advanced open optical fiber infrastructure. The group boasts an optical fiber network of over 61,700 kilometers and provides intelligent and open connectivity solutions to businesses, organizations, and public authorities. Eurofiber is supported by its shareholders Antin Infrastructure Partners, a French infrastructure-focused investment fund, and the PGGM Infrastructure Fund.

About Deepomatic

Deepomatic is a pioneer in the field of First Time Right Automation. Our AI-based computer vision platform is designed for companies operating critical infrastructures who strive to apply their quality standard at scale and increase their operations' First Time Right rate.

By analyzing photos taken by field workers and providing them instant feedback on the conformity of their job, our platform automates quality control. It guarantees the conformity of the data documenting daily field jobs and allows companies to follow the state of the infrastructure over time to make informed decisions to operate and maintain it.

Deepomatic serves leading international clients in sectors such as telecommunications, energy, hospitality and mobility. Nearly 30,000 field workers employed by Bouygues Telecom, Swisscom, CityFibre, Vodafone, Telefónica, Circet, and Unit-T use Deepomatic's solution every day, highly improving their productivity and efficiency on site. Our platform analyzes close to 2 million operations every month.