.png)
Vilans is very ambitious when it comes to data and AI, and would like to take steps with Large Language Models. Relevant guidelines, methods and tools in healthcare, for example, are still difficult for healthcare professionals to find and compare. They are spread across many different websites, not all of which are trustworthy.
Het resultaat
We developed a customized chatbot that helps healthcare professionals find information from reliable websites. The chatbot first helps the user clarify their question and then provides references to multiple useful sources. An LLM is used for various subtasks, but it is not the core of the chatbot, which further contributes to reliability.
The test with experts shows that the chatbot effectively helps clarify the question and finds relevant information in 90% of the conversations. All answers are completely reliable, no information is given that cannot be found in the sources.