Recommended workflow for biodiversity assessments with bulk samples and DNA metabarcoding that obtains count and biomass data with species level taxonomic identifications. A Specimens from a bulk sample (bottle) are first processed individually. B Processing includes specimen photography (camera), specimen counts (abacus), body size measurements (caliper) and biomass estimation (scales). Ideally, this is done automatically (green robot icon) and involves automatic image recognition to achieve preliminary taxa identifications on broad taxonomic scales. C Specimens are then re-combined to a community sample, a spike-in is added and DNA is extracted (microcentrifuge tube). D DNA metabarcoding delivers species level identifications and raw read counts (ASV table), which are corrected via the spike-in. E Image data is combined to a taxon list containing count, size and biomass data (taxon list). F Image data and DNA metabarcoding data are combined using machine learning approaches (data assembly, orange robot icon) to obtain a data set that contains information on species level identities, along with count data and biomass estimates (taxa bubbles), abbreviations: ASV – amplicon sequence variant, artwork: Alice Scherges.