High‐throughput DNA sequencing offers an efficient tool for assessing the taxonomic content of bulk arthropod samples. Many current DNA extraction protocols however require extensive handling of samples, like specimen‐based DNA extractions, or sorting of samples and are thus unsuitable for large scale studies. Furthermore, protocols often include homogenization and thus imply partial or complete destruction of the sample constituents. The aim of this study was therefore to investigate steps related both to sample pre‐processing and DNA extraction of unsorted bulk arthropod samples and explore possibilities for simplifying sample processing and thus increase sample handling efficiency without losing taxonomic information. Using mock bulk arthropod samples, we compare laboratory steps related to DNA extraction and semi‐automatic handling. Specifically, we (a) assess whether aliquots of digest buffer from bulk arthropod samples adequately describe the community composition; (b) compare a non‐destructive and a destructive DNA extraction method; (c) compare a phenol/chloroform inhibitor removal method with the exclusion of the same; and (d) compare manual purification to automated DNA purification on a QIAcube laboratory robot. Using DNA metabarcoding and mock bulk arthropod samples, we show that it is possible to efficiently process unsorted arthropod bulk samples with a non‐destructive DNA extraction approach. We found that homogenizing samples yielded more DNA but also generally produced more inconsistent results when compared to non‐destructive extraction. When assessing the recovered taxonomic content of samples (operational taxonomic units, OTUs), intact samples performed at least comparable to, if not better, than homogenized samples. Additionally, we show that sample processing can be further simplified from using a defined volume of digest, no phenol/chloroform purification and automated DNA purification. This approach can be a way to process hundreds of unsorted bulk samples effectively, consistently and with a minimum loss of valuable morphologic information.