Seasonal forecasts are generated with complex coupled atmosphere-ocean models. The weather parameters are expressed as a departure from climate values for one-month up to two years. Although seasonal forecasts are routinely produced at many operational and research centres, the reliability is still low over the most areas in the world. Improving the prediction skill is a very challenging task. DeepVisum produces special reports for specific clients based on seasonal forecast outputs, historical climate data and our AI technologies.