Three experiments were conducted to evaluate a synchronization protocol with AI at a predetermined time. In Exp. 1, 169 dairy heifers were assigned randomly to two groups: 1) timed AI (TAI), consisting of GnRH agonist injection (d 0, 1700), PGF2 alpha injection (d 7, 1700), GnRH agonist injection (d 8, 1700), and AI (d 9, 0800); and 2) AI at estrus (AIE), consisting of GnRH agonist injection (d 0, 1700), PGF2 alpha injection (d 7, 1700), and AI at detected estrus. Pregnancy rate was 25.8% for TAI (n = 89) compared with 48.7% for AIE (n = 80; P < .001). Experiment 2 was comparable to Exp. 1, but the second GnRH agonist injection in TAI was given 48 h after injection of PGF2 alpha. Heifers in TAI (n = 187) were inseminated at detected estrus if estrus occurred within 39 h after administration of PGF2 alpha (n = 47). Pregnancy rates were 45.5% for TAI and 48.0% for AIE (n = 177). Conception rate was reduced for TAI (45.5 [85/187] < 61.2% [85/139]; P < .005). In Exp. 3, the second injection of GnRH agonist, given at 48 h after injection of PGF2 alpha, was replaced with hCG (3,000 IU, i.m.). No differences in pregnancy rate were detected for TAI (52.9% [54/102]) vs AIE (56.1% [55/98]). Conception rate was reduced for TAI (52.9 [54/102] < 72.3% [55/76]; P < .005). Delaying the second GnRH agonist injection by 24 h improved pregnancy rate, but replacing the second injection of GnRH agonist with an injection of hCG did not prevent a reduction in conception rate.