The power of AI for summarizing meetings

We designed an AI-powered assistant to function as a powerful aid in certain situations

The problem & situation

You want to do something for the environment in your own home or you want to limit your carbon footprint when going on a holiday. However, you’re unsure where to begin and information seems to be different everywhere you look. Our client is a knowledge center that provides everything you need to know to get started with topcis such as these. Next to their interaction with consumers, they also have an advisory role towards governmental bodies and interest groups of certain communities.

To stay in touch with these various stakeholders and use their feedback for future research, knowledge sharing sessions are hosted on a regular basis. These sessions are information rich and can take up to several hours. To extract that information, they need to be carefully documented and summarised. A process that is currently managed by people, taking up a lot of time and resources.

The solution

To make this more efficient, we designed an AI-powered assistant to function as a powerful aid in these situations. The application uses audio from meetings and knowledge sharing sessions as input, to then automatically generate a summary. The summary contains general information from the session, but also takes into account detailed examples shared by the participants.

To use Generative AI on this case, we designed a Retrieval Augmented Generation (RAG) layer that successfully directs the Generative AI LLM model to find the necessary input before generating the summary. This approach ensures accuracy and keeps hallucinations at bay. Fun fact, while working with the AI-assistant, we were even able to take some of the human bias that wel all posses, out of the summary.

The outcome

By implementing the AI-assistant in the workflows, our client was able to greatly reduce cost and increase efficiency. In some cases, even improving on the quality of the summaries itself.