Individuals with Factor XI (FXI) deficiency have a variable bleeding tendency that does not correlate with FXI:C levels or genotype. Comparing a range of sample conditions, we tested whether the thrombin generation assay (TGA) could discriminate between control subjects (n = 50) and FXI-deficient individuals (n = 97), and between those with bleeding tendency (n = 50) and without (n = 24). The comparison used platelet-rich plasma (PRP) and platelet-poor plasma (PPP), either with or without corn trypsin inhibitor (CTI) to prevent contact activation, over a range of tissue factor (TF) concentrations. When contact activation was inhibited and platelets were absent, FXI:C levels did not correlate with thrombin generation parameters, and control and FXI-deficient individuals were not distinguished. In all other sample types, the best discrimination was obtained using TF 0.5 pM and assay measures: endogenous thrombin potential (ETP) and peak height. We showed that although a number of conditions could distinguish differences between the groups tested, TGA measured in PRP with CTI best differentiated between bleeders and nonbleeders. These measures provided high sensitivity and specificity (peak height receiver operating characteristic [ROC] area under the curve [AUC] = 0.9362; P < .0001) (ETP ROC AUC = 0.9362; P < .0001). We conclude that by using sample conditions directed to test specific pathways of FXI activation, the TGA can identify bleeding phenotype in FXI deficiency.