What is it about?
We created a web-based platform which makes it easy to test out chatbots, and rate them based on how well they perform. It allows you to have conversations with different bots, and judge which ones seem, for example, more human-like, accurate, and helpful in their responses.
Featured Image
Photo by Mohamed Nohassi on Unsplash
Why is it important?
Given the increase of chatbots in our everyday life, we need a way to rate them based on human interaction. Our platform makes it easy for anyone to evaluate chatbots through actual conversations, rather than just technical tests. By making the platform easy to set up, anyone can test their bots according to their personal questions and needs.
Perspectives
What made working on this article unique, is that it is not just a research article, but it is software that has been built from an actual need from within our research group. We do not only provide the platform for others to use, but actually use it in our yearly course ourselves, which is not often the case in our research work.
Daan van der Weijden
Universitat Zurich
Read the Original
This page is a summary of: Alan's Speakeasy - An Ecosystem for the Evaluation of Conversational Agents, May 2025, ACM (Association for Computing Machinery),
DOI: 10.1145/3701716.3715165.
You can read the full text:
Resources
Contributors
The following have contributed to this page
