Leveraging AI to transform energy needs into real-world solutions

Nearly 600 million people in sub-Saharan Africa still live without electricity. To overcome this challenge, new infrastructure is required, along with better-targeted investment, equipment maintenance, loss minimization and greater integration of renewable energy.

Artificial intelligence has a role to play in each of these stages. It can be used to analyze vast quantities of data, detect anomalies, anticipate future developments and make technical tools more accessible. It cannot replace investment or sector workers, but it can help them manage energy systems more effectively.

The Digital Energy program: moving from theory to practice

Since 2020, the Digital Energy program, funded by the European Union and implemented by AFD, has been accelerating Africa’s energy transition through digital innovation. The program has supported around 50 projects in over 15 countries by providing financing, technical assistance, training and networking opportunities for sector stakeholders.

The digital-energy.eu platform brings together around 2,000 partners, serving as a forum for sharing information on smart meters, geospatial data use, cybersecurity and mini-grids. In 2026, a training course was launched on generative and agentic AI to help companies – previous winners of the Digital Energy Challenge – develop their skills in this field, as well as to identify and test their own use cases.

“This approach is based on a simple principle: data first, AI second. Without reliable, accessible and organized data, even the most powerful tools can be difficult to exploit effectively. The challenge, then, is to identify a few high-impact use cases and integrate them into existing practices and systems, rather than running a vast number of standalone pilots,” says Cyril Renault, Digital Energy Program Manager at AFD.

Four companies trialing real-world solutions

This vision was put into action last July in Paris through an “AI Hackathon” organized by AFD and the Onepoint group. Selected from 13 applicants, four companies worked with experts to define a problem, identify a relevant use of AI and develop an initial prototype.

Fuel Switch is an automated system that verifies meter readings and produces the reports required for renewable energy projects looking to raise carbon financing. The goal is to provide reliable information, shorten processing times and facilitate access to project financing.

MyJouleBox, which operates mini-grids in West Africa, aims to use AI to resolve technical issues more efficiently. This could speed up troubleshooting for local mini-grid operators and reduce power outage times.

In Nigeria, ICE Solar is developing a chatbot that answers common queries from users of its solar equipment. It aims to provide guidance on installing or using equipment, resolve simple problems and refer more complex issues to the relevant teams.

Lastly, Roseau Technologies is exploring how AI could improve electricity grid modeling. An operator could define the scenario they would like to analyze (an electricity connection, an increase in demand, or the integration of a solar power plant, for example) and the software would then convert their query into technical parameters.

MyJouleBox crowned the winner

These four companies pitched their ideas to an audience of over 60 private and public sector stakeholders from Europe and Africa, including representatives from start-ups and utility companies. Each company left the AI Hackathon with a priority use case, prototype and roadmap. 

The French company MyJouleBox was crowned the winner by a jury made up of Antoine Djigbenou, Director of Digital Transformation (Eranove); Pierre Haessig, Assistant Professor (CentraleSupelec); Cyril Renault, Digital Energy Program Manager (AFD); and three AI and energy experts from the Onepoint Group, François Krempp, Florian Clarté and Tony El Hawid. MyJouleBox will receive 17 days of additional support from the Onepoint Group to help transform its prototype into a working solution.

This event exemplifies the Digital Energy program’s approach: take stakeholder needs as the starting point, assess the technology’s real value, test a solution, and then prepare for its roll-out.

AFD’s goal is not to promote artificial intelligence for its own sake, but rather to help companies and energy operators harness its potential to improve service quality, make infrastructure more reliable, and expand access to energy.