This paper proposes a forecasting strategy for a set of macroeconomic variables using information from surveys to professional analysts. Specifically, it is assumed that certain forecasts for the current state of the economy (nowcasts) are very difficult to beat in the short-term, so that there are benefits in including them in the time series of the variables to predict. For the U.S. economy, the Survey of Professional Forecasters (SPF) of the Federal Reserve Bank of Philadelphia is a renowned source of nowcasts and is therefore the starting point chosen to predict seven macroeconomic variables of interest. Using several models, both univariate and multivariate, it is possible to compare the forecasts that result from the use of this strategy with the predictions that would be obtained if the series did not include the additional information. Moreover, the performance of the models with nowcasts is compared with the predictions of the survey professional themselves. While the SPF asserts itself as highly reliable, the nowcasts appear to contribute for increasing the accuracy of the models used. Although sensitive to the choice of variables, the approach proposed in this paper proves to be quite promising and paves the way for further research, namely the application to other variables and/or economies.