Introduction Mycobacterium caprae is a member of the Mycobacterium tuberculosis complex (MTBC) not routinely identified to species level. It lacks specific clinical features of presentation and may therefore not be identified as the causative agent of tuberculosis. Use of whole genome sequencing (WGS) in the investigation of a family microepidemic of tuberculosis in Almería, Spain, unexpectedly identified the involvement of M. caprae. Aim We aimed to evaluate the presence of additional unidentified M. caprae cases and to determine the magnitude of this occurrence. Methods First-line characterisation of the MTBC isolates was done by MIRU-VNTR, followed by WGS. Human and animal M. caprae isolates were integrated in the analysis. Results A comprehensive One Health strategy allowed us to (i) detect other 11 M. caprae infections in humans in a period of 18 years, (ii) systematically analyse M. caprae infections on an epidemiologically related goat farm and (iii) geographically expand the study by including 16 M. caprae isolates from other provinces. Integrative genomic analysis of 41 human and animal M. caprae isolates showed a high diversity of strains. The animal isolates’ diversity was compatible with long-term infection, and close genomic relationships existed between isolates from goats on the farm and recent cases of M. caprae infection in humans. Discussion Zoonotic circulation of M. caprae strains had gone unnoticed for 18 years. Systematic characterisation of MTBC at species level and/or extended investigation of the possible sources of exposure in all tuberculosis cases would minimise the risk of overlooking similar zoonotic events.