Objectives were to evaluate the effects of implementing timed AI for first insemination on reproduction and cost per pregnancy in dairy heifers. Non-pregnant Holstein heifers (n = 611) at 400 d of age from 3 farms were allocated randomly to AI after detected estrus (Control, n = 306) or to timed AI for first AI (TAI, n = 305) followed by detection of estrus. Heifers in TAI were enrolled in the 5-d timed AI protocol on study d −6 [d −6, GnRH and an intravaginal progesterone insert; d −1, PGF_2α and insert removal; d 0, PGF_2α; d 2, GnRH + AI). Starting on study d 0, heifers had tailheads painted daily with chalk for detection of estrus. Control heifers received PGF_2α if not inseminated within 7 d of moving to the breeding pen. The study lasted 84 d to allow 4 21-d estrous cycle periods of breeding. A herd budget was created to determine cost per pregnancy. Data were analyzed by ANOVA, logistic regression, and the Cox’s proportional hazard model using the MIXED, GLIMMIX, and PHREG procedures of SAS and models included the effects of treatment, farm, and treatment by farm interaction. The interval to first AI was 8 d shorter (P < 0.01; 10 vs. 2 d) and the 21-d cycle insemination rate was greater (P < 0.01) for TAI than Control heifers (82.4 vs. 91.4%). Pregnancy at first AI (PAI; Control = 58.3% vs. TAI = 62.8%) and all AI (Control = 56.0 vs. TAI = 57.4%) did not differ (P > 0.15) between treatments. Implementing timed AI increased (P < 0.01) the 21-d cycle pregnancy rate from 47.9% in Controls to 57.2% in TAI heifers and reduced (P < 0.01) median d to pregnancy (Control = 14.0 vs. TAI = 2.0). In fact, the hazard of pregnancy improved (P < 0.01) 60% for TAI than Control heifers (adjusted HR = 1.60; 95% CI = 1.35–1.89), resulting in increased (P < 0.01) proportion of pregnant heifers by 84 d of breeding (Control = 85.2 vs. TAI = 91.5%). The cost per pregnancy for was $17.16 less expensive (P < 0.01) for TAI than Control heifers. Implementing timed AI for first service followed by detection of estrus improved reproductive performance and reduced the cost per pregnancy when compared with heifers managed under detected estrus only.