The presence of influential values in the estimation of population totals is an important problem that numerous polling organizations have to commonly deal with. While winsorisation methods seem to have become the most common technique to treat them, Beaumont et al. (2013) proposed to use the conditional bias to construct an equivalent robust estimator that lessens the effect of these outliers. C. Favre-Martinoz started to build the package REVIVALS (Robust Estimation for Variables with Influential VALues in Survey) with the idea of making these theoretical results available to the general R user. Resuming the work that he had been doing so far, we present several improvements to the package. These include some robustness enhancements in line with the theoretical conditions exposed in the literature, as well as some new features that facilitate its use. We also propose a real example application with some data about vacant homes collected in the Haute-Garonne French department. This illustrates the functions of REVIVALS and gives some lines of thought about the behaviour of the Horvitz-Thompson robust estimator in the presence of influential values.