Abstract

The increasing maturity of artificial intelligence technologies related to conversational systems, such as machine learning algorithms, natural language processing, linguistic data summarization, and natural language generation, is changing the way in which users interact with information systems. In this sense, two trends stand out: the improvement of rule-based smart chatbots and Large Language Models. These two models have different characteristics, with advantages and disadvantages for different scenarios. In this work, the BRasa ecosystem is proposed, which allows the construction of conversational systems that exploit the advantages of the two models. The ecosystem combines different computational intelligence techniques with natural language processing techniques, among which the following stand out: linguistic data summary, fuzzy logic, and neutrosophic theory. As part of the validation of the proposal, using the ecosystem, two conversational systems are built for different environments. The experimental design consists of different tests that demonstrate the applicability and effectiveness of the proposal.