ABSTRACT The systematic discovery of outflows in the optical spectra of low-mass X-ray binaries opened a new avenue for the study of the outburst evolution in these extreme systems. However, the efficient detection of such features in a continuously growing data base requires the development of new analysis techniques with a particular focus on scalability, adaptability, and automatization. In this pilot study, we explore the use of machine learning algorithms to perform the identification of outflows in spectral line profiles observed in the optical range. We train and test the classifier on a simulated data base constructed through a combination of disc emission line profiles and outflow signatures, emulating typical observations of low-mass X-ray binaries. The final, trained classifier is applied to two sets of spectra taken during two bright outbursts that were particularly well covered, those of V404 Cyg (2015) and MAXI J1820+070 (2018). The resulting classification gained by this novel approach is overall consistent with that obtained through traditional techniques, while simultaneously providing a number of key advantages over the latter, including the access to low-velocity outflows. This study sets the foundations for future studies on large samples of spectra from low-mass X-ray binaries and other compact binaries.