In recent years, dynamic conditional score-driven (DCS) models have attracted lot of interest in Economics, Finance, and Econometrics. However, their potential extends far beyond. The reason lies in the simplicity of the approach to time-series modelling and the easiness in parameter estimation. After presenting some motivating examples, in the first part of the talk I review the main theoretical properties of DCS models and discuss the flexibility of the approach in empirical applications. In the second part, I detail an application to high-frequency financial data. The approach has proved to be very effective in disentangling the fundamental price dynamics from micro-structure noise and in recovering the seasonal behaviour of prices at intra-day level.