The WEDO project, financed by InnoSuisse, aims to integrate advanced Large Language Model (LLM) tools into the WEDO applications.

WEDO, a Fribourg-based startup, is dedicated to enhancing workplace productivity through its meeting, task, and team management platform. Trusted by over 600 European companies, WEDO is expanding its offerings by embedding AI-driven tools into its application.

This project investigates the opportunities offered by LLMs for meeting tasks. Some research questions are also explored as: How to design an efficient prompt? How to evaluate the qualities of the output of text generation? Are Retrieval-Augmented Generation (RAG) techniques offered benefice compared to prompt only request? Which trade off between a large model with a certain cost vs a smaller model which is maybe less accurate?

Among the tasks we are trying to automatised to streamline team workflows we can cite generating meeting agendas, drafting minutes summaries, and transcribing and summarizing discussions. The integration of local, privacy-focused AI ensures that users benefit without compromising their data. All models used are deployed locally through Ollama, assuring security, budget efficiency, and confidentiality—core.

To read more about WEDO, Forbes wrote an article about them: Forbes Highlights WEDO.