PrecisionSolar

Our solution aims to help close the electricity access gap for the 600+ million people in developing economies. We address this challenge by supporting companies and organizations that deliver products that enable electrification — solar home systems, commercial & industrial (C&I) solar systems, and microgrids — to do so more efficiently. 

Currently, we think electrification is not efficient for two main reasons: the cost of electrification is high (part of the cost is customer acquisition cost), and electrification certainty is not clear — some areas are harder to electrify than others. 

Companies that provide solar products need to spend a significant amount of time and financial resources searching for customers that have the right capacity and willingness to afford their products and the right energy demand size. In the case of microgrid developers, significant cost and time commitment is required for site visits and ground surveys for multiple potential sites due to the lack of certainty. These barriers result in the slow progress of electrification, hindering electricity access for communities. 

PrecisionSolar is a software for companies and organizations working on electrifying communities through solar technologies. PrecisionSolar can identify and size areas with predicted energy demand, suggest the right technology to deploy given affordability factors, provide climate and financial risk assessment associated with target electrification areas, and guide the organization’s agents on optimal deployment routes given geographic features of the target areas. 

Solar companies first input a question into the software's chat box such as, "I'm looking to distribute solar lamps in Eastern Kenya. Where should I set up distribution centers?" The software then runs through an integrated database and generates optimized suggestions. 

PrecisionSolar integrates many data sources using machine learning. Some of the high-level information integrated includes predicted wealth indices on a sub-county level, types of crops harvested nearby and at what time of the year, access to infrastructure such as major road networks, and predicted electricity demand. We also layer in time and risk factors — helping organizations determine which areas make sense to electrify first and the types of risks associated with each area, including long-term climate-related risks. 

For residential and C&I solar companies, PrecisionSolar assists with generating high-quality leads and routes that maximize the number of potential clients who can afford solar products. For microgrid developers, PrecisionSolar helps sizing electricity demand in different potential locations and associated risks. 

Currently, PrecisionSolar works in Kenya and Malawi, although the underlying algorithms work across Sub-Saharan Africa and in other developing economies. In the future, we envision PrecisionSolar to work across geographies and be able to tailor to the needs of all types of organizations (such as retailers, developers, and planning agencies.) We also aim to expand PrecisionSolar into PrecisionEnergy, a tool that is technology agnostic, working across solar, wind, biogas, and other energy technologies that support electrification in developing economies. 

Our solution serves two main groups: companies and organizations that deliver solar products, and communities that benefit from solar electrification in developing economies. 

The first group is composed of small- and medium-sized residential solar retailers, C&I solar retailers, microgrid developers, and national electrification agencies. These organizations typically have teams of 10 to 100. Oftentimes, these companies and organizations can benefit significantly from the integration of technology in their processes. There are an estimated 50,000 such organizations across Sub-Saharan Africa. 

The second group is the communities that benefit from solar products, primarily in semi-urban and rural settings in developing economies. There are more than 600 million people that belong to these communities. As PrecisionSolar equips organizations to more efficiently deliver solar products at scale, it in turn makes more of such products become available to these communities, facilitating the access to clean electricity. 

Our co-founders met at Stanford as undergraduate students, started the organization from bullet points on a whiteboard, and pushed the growth of the organization with our passion and belief that everyone in the world deserves clean and affordable energy, regardless of their location and level of wealth. We are also driven by the idea that technology should be a force for good in the world. These two beliefs motivate us to work on PrecisionSolar and make it an accessible and useful tool for everyone in the solar community. 

Our team is composed of and advised by computer scientists, finance professionals, renewable energy developers and researchers, and entrepreneurs on the ground in Kenya, Malawi, Sierra Leone, and other countries. Our team collectively has graduate-level training in computer science, economics and finance, and energy engineering; we have work experience applying those skills at organizations such as Google, McKinsey, and the World Bank. Our team also has field experience in close proximity to the communities which we aim to serve: our team has worked in Kenya, Rwanda, and Sierra Leone for various organizations. Our corporate partners that advise us and share field data with us include companies in Togo, Tanzania, and the DRC; our academic partners include Stanford Seed, an initiative of the Stanford Graduate School of Business with a network of 100+ companies in developing economies. We constantly seek guidance from these partners and are grounded by their insights and knowledge as we develop PrecisionSolar. 

We selected the stage above because we are currently iterating the software and testing the solution with a few corporate partners. We plan to launch the pilot version by early summer. 

We are applying to MIT Solve because we resonate with the belief that technology and social innovation can do substantial good in the world. We would like to seek guidance from the Solve community on organization-building in order to prepare for the next leg of growth. 

As PrecisionSolar pilots, our top-of-mind question is how to improve and scale the product, how to refine our value proposition, and how to build a strong team. We would benefit from technical assistance as well as guidance on organization-building: we think the Solver community is a great resource, where we can learn from other nonprofits that have gone through different phases of growth and faced different challenges. We hope to be long-term contributing members of the community and help others where we can as well. 

Our solution is innovative in 2 ways. First, it is tailored for the solar company/organization to use. The user can simply input text prompts much like ChatGPT and then receive optimized suggestions in terms of solar electrification. We are not aware of other software products with this feature. Second, we aggregate a wide array of data to build the PrecisionSolar engine, which exceeds the capabilities of many existing solutions in the market. Third, we use machine learning and computer vision algorithms to integrate data with satellite images, an approach that can enhance the usefulness of the tool for solar companies. 

Our theory of change is that increasing electricity access is fundamentally a matching problem — matching available, affordable supply with right-sized demand. If we focus only on entrepreneurial innovation, we think enabling small- and medium-sized enterprises to provide the right products to the right locations at higher efficiency with the help of technology will reduce the inefficiencies in the matching process. If SMEs can better deliver electrification products such as solar home systems and microgrids with more speed and precision, then communities and the SMEs both benefit, while electricity access improves. 

Our impact is measured through two indicators: the number of communities and households that are electrified because of products delivered with the help of PrecisionSola, and the additional revenue created for small- and medium-sized enterprises given the integration of PrecisionSolar into their deployment processes.

We run satellite images through computer vision algorithms to derive primary insights, such as detecting which regions are currently electrified using presence of nightlight, population density based on building outlines, and suitability of solar energy based on solar irradiance. We also leverage large-scale pre-training to predict socioeconomic factors and energy demand from satellite images and from other available data sources, even in places that have very little labeled data. 

We then use AI to integrate all of the above and to make it easier for our users to interact with our platform in natural language, and show our final product as an interactive web-based map. 

We intend to use a subscription-based, tiered pricing approach targeting solar companies. The basic version is free and allows the user to access the map as well as basic information about the settlements. We charge a flat fee for companies to unlock the plus version and other premium versions that showcase more dynamic, in-depth information.