AbstractThere is a constant need to reduce turn-around times and keep costs as low as possible for the carriage screening of GBS in pregnant patients. Laboratory automation might provide an edge in this field. The objectives of the present study were: i) to compare the performance of the direct chromID™ Strepto B agar (CA) plating against LIM-broth enriched plating on CA for the detection of GBS from vagino-rectal screening-swabs; and ii) to assess the usage of PhenoMATRIX™ for the automated screening of GBS. Between January 2021 and December 2023, 9′107 vagino-rectal specimens were collected from pregnant women at Geneva University Hospitals and were used to address the first objective. There was a small difference in the GBS detection rates between direct CA plating (13.2%; 1′202/9′107) and LIM-broth enriched plating on CA (13.2%; 1′198/9′107). Based on the LIM-broth enrichment results, the sensitivity and specificity of the direct CA plating were 98.3% (95% CI, 97.3%—98.9%) and 99.7% (95% CI, 99.5%—99.8%), respectively. Importantly, for 25 specimens, GBS growth was only detected by direct CA plating. We used a random set of 8′768 CA plate pictures for the machine learning of PhenoMATRIX™. The validation was carried out on an additional set of 830 CA plate pictures. The sensitivity and specificity of PhenoMATRIX™ were 100% (95% CI, 96.6%—100.0%) and 90.2% (95% CI, 87.8%—92.1%), respectively. We established that for GBS screening, the performance of direct CA plating is not inferior to the LIM-broth enriched approach. By relying on PhenoMATRIX™, the negative predictive value for GBS screening reaches 100% (95% CI, 99.4%—100.0%), enabling the automatic release of GBS-negative cases within 24 h.