The aim of the study was to clarify the phylogenetic relationships among Northern European Ancistrocerus and comparison of the applicability of evolutionarily neutral and non‐neutral markers for reconstruction of phylogeny. We used a 19,400 bp long dataset that included parts of mitochondrial DNA, nuclear rDNA operon, and 10 nuclear protein‐coding genes. Application of molecular barcoding and species delimitation algorithms unveiled a presence of cryptic species, A. balticus sp. n., in the trap‐nesting wasp communities of the centre of Europe. We assessed the morphological, biological, and ecological differences of it from the sibling A. trifasciatus and updated the regional identification key. Phylogeny reconstruction using the neutral and the presumably non‐neutral markers resulted in different tree topologies. Evolutionary congruence of the rDNA operon with the other markers was relatively low. Evolutionary rate of the mitochondrial genes was 7–8 times as high as that of the exons of the nuclear genes, therefore, the mitochondrial markers overshadowed the nuclear ones in the phylogeny reconstructions. We assumed that at the speciation level, we might consider two different patterns of phylogeny: one based on evolutionary time and neutral changes, and the other based on adaptive evolutionary pathways under directional selection pressures. We assessed the effect of directional selection on the nuclear protein‐coding genes, applying the Spearman's rank correlation between pairwise phylogenetic distances among species, estimated using exons, and these distances, estimated using introns. One of these markers demonstrated a lack of positive correlation, implying a variable directional selection pressure on the coded protein. The publication has been registered on ZooBank: urn:lsid:zoobank.org:pub:13BD28D0‐736D‐4B2A‐B5CF‐4824BD4CDCFB.