Bridging the digital divide: How Vua is bringing GenAI and SMS to small-scale fishers

Co-authored by: Carmen Merab and Juan Diego Martin Soto

August 15, 2024 - 2 mins read

Startups are increasingly using Generative AI (GenAI) to streamline their operations and reach more customers. This is becoming more prevalent in developing countries, where entrepreneurs are finding ways to simplify complex technologies and adapt them to local contexts, such as automating tasks, personalizing experiences, and providing valuable insights for diverse business contexts. The game-changer is when startups combine these technologies to reach users who may not have access to smartphones or reliable internet connections.

One such startup is Vua, a TECA alumni company providing financial services to small-scale fishers. In 2023, they received a grant from Climate Collective to leverage AI, and other digital infrastructure to scale nature and climate action as part of a product hatchery that provides critical support to entrepreneurs who have completed the TECA program, making grant capital available at the proof-of-concept stage.

Recognizing the critical role of SMS in providing financial services to small-scale fishers, Vua used the grant to address the challenges small-scale fishers face in accessing financial services and knowledge bases on feature phones.¹The Vua team in collaboration with the TECA team, identified how SMS integration with GenAI could help these communities increase their financial literacy. Unlike other workers, small-scale fisherfolk primarily rely on feature phones due to their affordability and durability. This is crucial in their work environment, where smartphones could be easily damaged on boats.

Vua developed a system that could enhance the information retrieved from OpenAI’s Large Language Models by ensuring that it references an authoritative knowledge base to use the most recent and trustworthy data available on the topic. In technical terms, the system Vua developed to do this is called a RAG (retrieval-augmented generation), and Vua made all this available via SMS.

Once the system was developed and its code tested and validated, Vua deployed it among 100 users on the north coast of Kenya and monitored the system’s performance and user engagement with it.

Given that other companies in the climate space in Africa might face challenges similar to those encountered by Vua users, both teams agreed that it would be beneficial to open-source the Vua code and implementation documentation to help other companies leverage the same resources.

The details of the code and the documentation are available in Python to allow fast prototyping and in Java to allow robust implementation in production-grade environments. A step-by-step guide has also been provided to configure AfricasTalking, enabling the system to deliver content via SMS to 27 African countries.

The use of this technology goes far beyond fisheries and can be used in other sectors for:

Enhanced digital customer support via SMS such as offering crop improvement tips based on specific varieties and locations.
Personalized climate-related emergency advice such as which essential goods to prioritize, how to prepare homes to minimize damage, and how to pack an emergency kit.
Customized learning journeys to increase climate adaptation awareness and preparation, including information on financial products like insurance and loans.

And many other use cases entrepreneurs in Africa will think of.