We are excited to be developing and testing together with IRCWASH WASH Systems Academy, an AI Learning Companion to help onboard and support WASH professionals to learn about WASH systems and the building blocks. IRCWASH and WASHNote signed a Memorandum of Understanding at the start of 2024 to make this a reality.

It is built from two components:

  1. The Knowledge Companion, that provides a validated set of information and a model of the conceptual framework to be used.
  2. The Learning Companion, a conversational agent that can chat with users via mobile apps such as Telegram, Whatsapp, Signal or Element.

The Learning Companion document  and data repository focuses on WASH systems and their building blocks, and the conceptual framework of the WASH Systems Academy courses (IRC building blocks).

These components are easily extensible and it is possible to switch conceptual frameworks (knowledge graphs) on the fly to be able to map, for example, results from one data source to another methodology. Ultimately, these will be used to help in monitoring and evaluation and to support the development of validated public datasets.

So far, together with a small team, we’ve built a knowledge graph of the key concepts used, a document repository that can be easily cited by the learning companion and a Telegram bot. Much of our effort has been used to build up our user journeys, validate with existing users and develop a evaluation framework for both improving conversations and evaluating the outcomes of the exchanges.

We’ve also researched how to run this AI product in an affordable manner so that it can scale and achieve an impact on WASH services, which is the ultimate goal. Some of this research will be published shortly.

In the coming weeks, we will be testing with users in Rwanda and Ethiopia and we look forward to sharing our findings. As the product matures, the relevant software components will be open-sourced and shared online. More importantly, we are building with open source and out of principle also contribute back our improvements to these original open-source projects so that the bigger community online can benefit.