Performing pregnancy diagnosis sooner after TAI could decrease days open in dairy cows if non-pregnant cows are enrolled in resynchronization programs. Methods for accurate pregnancy diagnosis within 3 wk after TAI are based on interferon tau-stimulated gene (ISG) expression in blood but milk samples are often more convenient to obtain than blood samples. The objective was to assess the utility of measuring ISG15 expression in milk somatic cells as a method to diagnose pregnancy in cows after TAI. Blood (10 mL) and composite milk (200 mL) samples were collected from 48 primiparous and 13 multiparous Holstein cows (n = 61; 102 ± 12 DIM; 36 ± 20kg/d) at 18, 20, and 22 d after TAI. Samples were placed on ice after collection and RNA was extracted on the same day. RNA samples were used for cDNA synthesis and cDNA was used in RT-PCR analysis of gene expression for ISG15 and cyclophilin (reference gene). Ratios of ISG15 to cyclophilin (ICR) were calculated. Milk somatic cell expression ratios were log-transformed before analysis to reduce variance. Transrectal ultrasonography diagnosis for pregnancy at 33 d or 35 d after TAI was the reference standard. The ICR of blood and milk cells were tested for the effects of pregnancy status, day, parity, and interactions using the MIXED procedure of SAS 9.4 (Cary, NC). The REG procedure of SAS was used to determine the correlation between milk and blood ICRs. The ICR was greater in blood of pregnant cows (0.58 ± 0.07; n = 28 compared with non-pregnant cows (0.12 ± 0.06; n = 33) on d18, 20, and 22 (P < 0.0001). In same cows and on the same days, milk somatic cell ISG15 expression was also greater in pregnant (0.64 ± 0.17) compared with non-pregnant (0.20 ± 0.16) cows (P < 0.059). Day of sampling did not affect ISG15 expression for either sample type. Overall, ISG15 expression in both blood and milk somatic cells was greater for pregnant compared with non-pregnant Holstein cows. Testing milk for ISG15 expression may be an alternative to ISG15 testing in blood.