<p>Insect herbivory constrains global crop productivity and reliance on chemical insecticides is increasingly untenable given resistance evolution, environmental costs and tightening regulation. This review argues that the plant holobiont the plant together with its rhizosphere, phyllosphere and endophytic microbiota offers a more complete account of herbivore resistance than plant-centred models alone, but that the field's central problem is no longer discovering mechanisms; it is explaining why those mechanisms perform inconsistently outside controlled conditions. We synthesize evidence for three principal routes by which microbiomes shape herbivore outcomes direct production of insecticidal metabolites, priming of jasmonic acid (JA), salicylic acid (SA) and ethylene (ET)-mediated defences and disruption of herbivore gut microbiota and show that each is mechanistically well supported yet fragile in the field, where genotype, soil chemistry, climate variability and agricultural management interact non-additively to determine outcomes. Multi-omics and machine-learning approaches have, in an important sense, outpaced this translational challenge, they can identify resistance-associated taxa and predict community function with increasing precision, but that predictive power has not yet translated into consistent field performance, stable formulations or regulatory clarity. We evaluate where the evidence for microbiome-mediated resistance is strong versus preliminary, reconcile several apparently contradictory findings on environmental modulation and argue that closing the lab-to-field gap rather than further mechanistic discovery is the binding constraint on using the plant holobiont as a pillar of climate-resilient integrated pest management.</p>