Bees are a diverse group of flower-visiting insects with over 20,000 described species. Understanding their phylogenetic relationships is essential for studying their biology and natural history. A central theme of this dissertation is the use of ultraconserved elements (UCEs) to reconstruct phylogenetic relationships. UCEs are widely shared, highly conserved DNA sequences that can be efficiently obtained using high-throughput DNA sequencing platforms. I use UCEs to shed light on the relationships of the corbiculate bees and show that base compositional biases of DNA sequences are a pervasive source of gene tree and species tree incongruence in this group. Accounting for these biases reveals that the Neotropical orchid bees are the earliest branching lineage of this group, a finding which contests previous results of smaller molecular data sets and morphological data. In another study, by searching for UCEs or similar sequences in published genomes, I found that some UCEs are incredibly widely shared among arthropods and beyond, and that even humans, fungi and bacteria can share UCEs that have very similar sequences to that of insects. I demonstrate that UCE sequence data can be combined with data obtained through transcriptome sequencing. Using this combined approach, I reconstruct the phylogeny of Apidae and establish a revised, rank-based taxonomy of the group. Next, I explore estimation error of gene trees from UCE sequence data and compare different approaches for minimizing the introduction of noise for gene tree summary methods. Using this data, I establish a molecular phylogeny of the Pseudapis group, a lineage of nomiine sweat bees endemic to the Old World. Lastly, I describe three new bee species from Africa: a new species of the previously monotypic Schwarzia, and two new species of Pseudapis.