PurposeThis research aims to address the transformative service research (TSR) agenda by examining the issue of caste-based financial exclusion in microcredit lending services in India. To do so, it draws on statistical discrimination theory from labour economics to develop and test a multi-level prosocial service orientation framework.Design/methodology/approachSurvey data come from 238 loan officers and 250 lower caste loan applicants across 43 microfinance institutions (MFIs) in India. The data are analysed using hierarchical linear modelling, a method appropriate for investigating micro- and macro-level organisational variables.FindingsAt the micro level, the service orientation factors of social dominance orientation and algorithmic-driven lending decisions affect financial exclusion of lower caste bottom-of-the-pyramid (BoP) vendors. At the macro level, the service orientation mechanism of inclusive service climate reduces caste-based financial exclusion, while the level of lending risk to reduce discrimination receives no support.Research limitations/implicationsResearch in other contexts is warranted to confirm the prosocial service orientation model. Methodological challenges at the BoP also present avenues for insightful work.Social implicationsThe study shows the importance of an inclusive service climate and reassessment of algorithmic-driven lending decisions to eliminate caste-based indicators in lending decisions. It also recommends policy reform of caste-based affirmative action at the macro- and micro-levels of lending decisions.Originality/valueThis research extends the TSR agenda to include caste-based discrimination in prosocial services. It takes a multidisciplinary perspective on services research by incorporating statistical discrimination theory from labour economics to extend understanding of service orientation.