Current methods for antimicrobial susceptibility testing (AST) are too slow to affect initial treatment decisions in the early stages of sepsis, when the prescriber is most concerned to select effective therapy immediately, rather than finding out what will not work 1 or 2 days later. There is a clear need for much faster differentiation between viral and bacterial infection, and AST, linked to earlier aetiological diagnosis, without sacrificing either the accuracy of quantitative AST or the low cost of qualitative AST. Truly rapid AST methods are eagerly awaited, and there are several candidate technologies that aim to improve the targeting of our limited stock of effective antimicrobial agents. However, none of these technologies are approaching the point of care and nor can they be described as truly culture-independent diagnostic tests. Rapid chemical and genomic methods of resistance detection are not yet reliable predictors of antimicrobial susceptibility and often rely on prior bacterial isolation. In order to resolve the trade-off between diagnostic confidence and therapeutic efficacy in increasingly antimicrobial-resistant sepsis, we propose a series of three linked decision milestones: initial clinical assessment (e.g. qSOFA score) within 10 min, initial laboratory tests and presumptive antimicrobial therapy within 1 h, and definitive AST with corresponding antimicrobial amendment within an 8 h window (i.e. the same working day). Truly rapid AST methods therefore must be integrated into the clinical laboratory workflow to ensure maximum impact on clinical outcomes of sepsis, and diagnostic and antimicrobial stewardship. The requisite series of development stages come with a substantial regulatory burden that hinders the translation of innovation into practice. The regulatory hurdles for the adoption of rapid AST technology emphasize technical accuracy, but progress will also rely on the effect rapid AST has on prescribing behaviour by physicians managing the care of patients with sepsis. Early adopters in well-equipped teaching centres in close proximity to large clinical laboratories are likely to be early beneficiaries of rapid AST, while simplified and lower-cost technology is needed to support poorly resourced hospitals in developing countries, with their higher burden of AMR. If we really want the clinical laboratory to deliver a specific, same-day diagnosis underpinned by definitive AST results, we are going to have to advocate more effectively for the clinical benefits of bacterial detection and susceptibility testing at critical decision points in the sepsis management pathway.