Awareness via chatbot: Collaboration with ReSViNET
Discover how we partnered with ReSViNET to build an AI-powered chatbot that provides global RSV education, available via web and WhatsApp.
Over the past few months I have been working on a project that has been both special and challenging: developing a chatbot for ReSViNET. For those who are not familiar with this organisation: ReSViNET is committed to reducing the impact of the Respiratory Syncytial Virus (RSV).
This virus causes severe pneumonia and can be fatal, especially for vulnerable groups such as infants, the elderly and people with underlying conditions. The impact can be enormous, especially in areas where medical care is less advanced.
The Origin of the Project and Technical Innovations
ReSViNET approached me with the question if I could develop an AI-driven chatbot that would provide education about RSV. The goal was to inform users about infection prevention, the latest developments in treatments and upcoming vaccines.
We decided to collaborate with doctors from the UMC Utrecht for this, which resulted in an extensive knowledge base. In addition, we use the technology and expertise of Enreach Dialox to realize the chatbot and the linked channels.
What made this project really special was the use of Retrieval Augmented Generation (RAG) . This means that we don’t just rely on AI models, but also carefully extract information from our knowledge base to provide users with accurate and detailed answers. When we started, we were still working with GPT-3.5, but we have since upgraded to GPT-4o mini, which resulted in even better results.
Multilingual chatbot
One of the unique aspects of the chatbot is that it is available in multiple languages, thanks to the language capabilities of OpenAI’s models. This allows us to truly serve a global audience.
This is where AI is particularly powerful: until now, it has certainly not been possible to provide personalized information at this scale in so many languages. It is a unique opportunity for companies and organizations that operate globally.
Multichannel Approach: From Web to WhatsApp
In order to reach as many people as possible, we decided to make the chatbot available not only through the ReSViNET website, but also through WhatsApp . This turned out to be a smart move, especially in countries where many people have skipped the step to computers and are using mobile devices directly. There are about 3 billion WhatsApp users worldwide, which gives us a huge reach.
A great thing about WhatsApp is that users can continue their conversations. Imagine someone asks a question, gets an answer, and then comes back a week later with a follow-up question—the chatbot remembers the conversation.
Especially when people are confronted with the disease, for example with their child, the world is in disarray and there are many concerns but also developments. It is precisely at such moments that there is a need for a developed conversation that can be approached practically. This creates an ongoing dialogue that adds a lot of value to the user experience.
QR-codes
In addition, we consciously use QR codes to facilitate the transition from the physical to the digital world. Imagine a hospital in Indonesia hanging up a QR code that allows patients to immediately start a conversation with the chatbot to get answers to their questions. This not only makes access to information accessible, but also easy to distribute.
The Learning Chatbot
Although the chatbot can already answer many questions, this remains a continuous learning process. We monitor the interactions to see where the chatbot falls short and add new information when necessary. We also expect developments regarding the availability of vaccines and expect questions about this.
The knowledge base offers the possibility to be up-to-date, actually much more than the general AI language models . The great thing is that we don't know exactly what questions users will ask in the future, but by continuously learning and improving, we can increasingly respond to their needs.
Experience It Yourself!
Are you curious? I invite you to have a conversation with the chatbot yourself via the ReSViNET website or via WhatsApp. It is a unique opportunity to experience how AI and human insight come together to make a difference in healthcare.
Rob Hoeijmakers
I will write more about this in the coming months and keep you updated on the effect. Sign up for the newsletter to stay updated.
