Towards building intelligent speech interfaces through the use of more flexible, robust and natural dialogue management solutions

► We tackle the problem of identifying the best practices when designing and evaluating a spoken dialogue system. ► We measure the relevance of carefully designed contextual information handling strategies. ► An evaluation is carried out in terms of both subjective and objective metrics including a thorough and comprehensive correlation analysis among them. ► A more natural, flexible and robust dialogue is possible through a Bayesian Networks-based solution for dialogue modelling. ► Results show that the intelligence of our speech interface has been well perceived highlighting its excellent ease of use.