Integrative taxonomy has emerged as a methodologically sound method for discovering new taxa. By coupling different sources of information (e.g. molecular and morphological data) and by combining analytical approaches (e.g. multivariate analyses, phylogenetics), systematists have the tools to uncover cryptic diversity and draw more reliable species hypotheses. Integratively-delimited taxa may be more prone to formal description compared to taxa delimited with a single source of information, because more detailed species hypotheses can be drawn. Integrative taxonomy has therefore the potential to reduce the so-called “taxonomic impediment” (i.e. the gap between the number of taxa that remain to be described and the number of taxa that are formally named). However, bridging fields such as molecular systematics and morphological taxonomy may remain challenging, as it either requires the training of versatile scientists, or the establishment of collaborations among scientists from these fields. We therefore asked ourselves whether integratively-delimiting species necessarily led to formal taxonomic descriptions. We reviewed the “integrative taxonomy” literature of the past five years to estimate the ratio of studies delimiting species (without describing them) over studies describing species, and tested whether this ratio was dependent on the types and numbers of characters used for delimitation. Based on a sample of 361 studies claiming an integrative approach to species delimitation, 68% used at least two characters for species delimitation, and 32% were based on a single character (most often DNA sequences), restricting their integrative approach to including different types of data analyses. Overall, 46% of studies delimit without describing, and a majority of those do not provide a justification for the absence of formal descriptions. Authors that do provide a justification mention lack of data, forthcoming publications with formal descriptions, or find nomenclatural rules too complex. Integrative taxonomy is, based on our sample, not an efficient solution to the “taxonomic impediment.”