Top: separation of SNIa and SNIbc transients into three groups (I, II, III) based on the SNIa probability reported by the FollowMe classifier. G-I with P(SNIa)>0.95 encompasses most SNIa events, while G-III captures almost all SNIbc but with a small SNIa contamination. G-II highlights the interesting area where the classifier does not confidently identify a transient as SNIa or SNIbc. Bottom: confusion matrix for SNIa and SNIbc transients in G-II, based on the most probably transient kind reported. Also within this group, most transients are correctly classified. The residual uncertainty expressed here can be taken into account by statistical models for cosmological analysis, but an observational campaign would be needed to verify whether the reported uncertainty is correct.