Using Predictive Analytics to Promote Off-grid Solar Connectivity in Kenya: Atlas AI and Engie Energy Access Partnership Case Study

The off-grid solar home appliance market is expected to electrify 624 million people by 2030, and is a key component of SDG7, which is the Sustainable Development Goal targeting to "ensure access to affordable, reliable, sustainable and modern energy for all."

Engie Energy Access is a global leader in off-grid clean energy solutions, and was looking for ways to accelerate the adoption of Engie solar appliances across the Kenyan market through a data-driven approach. To achieve this, Engie turned to Atlas AI, a predictive analytics company that uses high-resolution geospatial data and machine learning to provide location-based insights through its proprietary Aperture® platform.

The Problem

In late 2022, the Engie EA team identified customer qualification and acquisition as key priorities to drive the growth of its Kenyan business. Despite the company's reputation for product quality and year-over-year sales growth, it was still an early entrant to the Kenyan market and needed to find ways to connect with the highest potential customers in the markets with the most significant demand potential.

Engie EA’s customer acquisition strategy actually had three interconnected attributes: Identify and target potential customers in high-density areas of Kenya with inconsistent access to the grid, who can afford to repay consistently, and would be interested in purchasing Engie EA's off-grid solar home appliances.

Atlas AI’s data and platform proved to be an ideal solution for Engie to identify and monitor the local communities in Kenya where they could focus their growth efforts.

Questions Engie EA were seeking to answer:

Atlas AI’s Aperture Platform addressed these questions by:

Evaluating Engie sales performance in existing  markets to identify  market-related factors that influence strong sales performance.Recommending markets that are similar to existing high-performance areas and predicting expansion priorities and market potential.Learning from Engie’s performance over time to continue to refine predictions and adapt to changing market conditions and Engie’s evolving market footprint, sales performance and product mix.

About Atlas AI’s Aperture® Demand Intelligence Platform

Aperture is a cloud-based SaaS application running end-to-end on Google Cloud that blends Atlas AI's proprietary machine learning data with customer data to provide businesses with predictive business insights. The platform, powered by Google Earth Engine and BigQuery, offers customized market segmentation based on the unique operating context of each customer and forecasts key business outcomes on an ongoing basis, such as revenue potential and climate vulnerability, at a localized level, guiding operational and investment focus over time.

The platform enables both geospatial visualization and real-time business intelligence (BI) dashboards. The same data and insights can also be made available through existing internal BI and ERP tools such as Looker, SAP and Tableau through API integrations.

Engie deployment of Aperture
® Demand Intelligence, showing customer locations, sales hubs ands sales markets.


Atlas AI leveraged its location data, spending power, area with grid/connectivity, electrical access, population density data, and Engie's experiential market data to make informed recommendations.

Aperture® Demand Intelligence workflow.

What makes a high-performing ward in Kenya?

Using Aperture, Engie was able to identify the top-performing wards by sales, and found that to be considered high-performing a ward must have:

Aperture was able to filter and show wards that met those criteria and direct the commercial team on an ongoing basis to maximize sales potential.

Ward criteria examination within Aperture
® Demand Intelligence.

Market Prediction using Machine Learning

Engie harnessed Atlas AI’s machine learning-powered market performance analytics to answer the question “how many customers are likely to be in location X for product Y?”  The predictive model underpinning these insights takes into account dozens of demographic, infrastructure & historic performance variables developed by Atlas AI, and available exclusively via Aperture.

This model can be used to evaluate where in existing markets to continue investing, and where and when to expand to new territories.

This analysis for Bungoma County shows the expected penetration rate by Ward if hubs were opened across the county.

After analyzing the data, Atlas AI suggested the specific locations where Engie should concentrate its sales and expansion teams.

Field Validation

Atlas and Engie tested the Aperture performance models in the field to evaluate their reliability in relation to ground conditions. Working closely with the Engie Commercial team, Atlas AI identified areas of interest in three regions (Western, Eastern, and Coast), to target based on the selected criteria.

The hypothesis was that Engie concentrating sales activity in wards predicted by Aperture to be “high performing” would result in increased sales performance.

The Aperture® Demand Intelligence platform dashboard was configured for Engie to monitor sales and operations performance in POC territories on a daily basis.


To validate the performance predictions, Atlas AI set up a pilot program with Engie's commercial teams in three different regions in Kenya Western, Eastern, and Coast, with the first two being control groups. The Coast region's regional manager fully utilized Aperture® recommendations to target specific areas, leading to a 48% increase in monthly sales. In contrast, the Western and Eastern teams’ sales during the same time frame followed historical trends.

Atlas AI's Aperture® platform enabled Engie to identify the optimal regions for expansion in Kenya. Engie's sales teams leveraged this information, and the Coast region's performance demonstrated the effectiveness of the Aperture recommendation model.

The Future

A key benefit of Aperture® is the platform’s scalability across a wide range of markets, giving Engie the option to expand usage of the platform across their full operating footprint in a matter of weeks. More advanced customer segmentation analytics are also on the roadmap, supporting modeling of target customers by product type. This approach can enable Engie to lower customer default rates by avoiding overselling, a common challenge across the industry.


High-resolution market segmentation could be the key to unlocking a USD 100 million+ annual solar appliance business in Africa for Engie EA. By targeting specific high-density areas with unreliable grid access and identifying potential customers who have the income to repay consistently, the company could gain a competitive advantage over more established players in the market. Through hyperlocal market segmentation, Engie EA can create targeted marketing campaigns and tailor its product offerings to the needs and preferences of specific segments, thus winning competitive market share and achieving its ambitious growth goals.


Subscribe to our newsletter to get the latest news and updates.