LaPraS: Language Profiling and Segmentation

Although chat and voice bots operate as additional web care channels, they fall short in one essential area: their conversational capacities are limited. Usually, they produce standard responses, which sound robotic and distant. Chatbot users have indicated in multiple surveys that they miss the naturalness of a human conversation.

We're developing an algorithm that responds appropriately and empathically to Flemish chat and voice bots users between 18 and 24 years. After the implementation of the algorithm, the bots will be able to read and hear what emotion a consumer expresses and what personality traits s/he has.

This algorithm is useful for companies that rely on chat and voice bots as part of their web care strategy for young people. Today, 25% of Belgian companies are already investing in AI, and by 2021 that percentage will be 45%. The outcome of our project plays a crucial role in the reputation management of these companies: the more humanlike the online contact, the more positive the word-of-mouth.