Host attachment is often a critical step in the onset of pathogenesis. To attach to host cells, bacteria have evolved a range of adhesins that bind to specific receptors. Some of these adhesins have been thoroughly characterized using biochemical techniques. However, how adhesins engage with their receptors in a realistic context of host colonization remains obscure. For instance, how target cell surface properties regulate attachment has been overlooked. This hinders our understanding of pathogenicity, thereby limiting our ability to develop new therapeutic approaches. Here, we aimed at characterizing the biophysical rules underlying bacterial attachment to live cells. In this context, we displayed synthetic adhesins on both bacterial and mammalian target cell surfaces to study how the mammalian membrane microenvironment regulates attachment. By leveraging microfluidics and high-temporal resolution confocal microscopy, we tracked the early adhesion of bacteria to target cells and compared it to abiotic surfaces. We modeled the distribution of residence times and uncovered that the binding to mammalian cells is a two-step process, as opposed to one-step binding to an abiotic surface. In particular, we highlight the impact of the mammalian cell glycocalyx and of the actin-mediated cell remodeling. Altogether, our results demonstrate that adhesin-ligand binding is not the only regulator of bacterial adhesion, due to the host mechanical microenvironment playing a critical role on the initiation of infection. With a better knowledge of in vivo adhesion, we repurposed the synthetic adhesin system as a tool for bacterial-based therapy. Our plan consists in using adhesion to rewire host-pathogen interactions. By analogy with pathogenic viruses transformed into therapeutic gene delivery vectors, we focused on pathogenic bacteria injecting DNA to eukaryotic cells. Agrobacterium tumefaciens is a pathogen that delivers DNA to plants using its type IV secretion. It is widely used for gene editing in plants, and sometimes in yeast and fungal cells. It is therefore an attractive candidate as a human gene delivery vector. However, some cell types such as plant monocots or animal cells show extremely low transformation efficiency. Studies demonstrated a positive correlation between adhesion to recalcitrant plants and transformation efficiency. Hence, would a synthetic binding of A. tumefaciens to non-natural target cells increase delivery? To measure the impact of adhesion on delivery efficiency, we repurposed an endogenous autotransporter of A. tumefaciens to display the previously characterized synthetic adhesin. This significantly increased the binding to yeast and mammalian cells displaying the target surface receptor. In addition, we developed a split luciferase assay to quantify the transfer of helper proteins to target mammalian cell. This allowed us to optimize A. tumefaciens-mediated delivery to mammalian cells and to refine hypotheses concerning the translocation mechanisms involved in mammalian cells. Altogether, we show that synthetic adhesins are a valuable tool to improve our understanding of host-microbe interactions and for repurposing pathogens into therapeutic tools.