Automatic or robotic milking systems (AMS) are being adopted by dairy producers in the Upper Midwest United States at a relatively fast rate. There is limited research available on AMS in US dairies. This ongoing study is summarizing data collected from 52 AMS dairy farms in Minnesota and Wisconsin to describe characteristics of AMS systems and investigate factors that may influence AMS efficiency in US farms. Farms were visited once to collect housing and management information and then daily data from the AMS were remotely collected for approximately 18 mo. These data showed that on average cows were milked 2.61 ± 0.30(mean ± SD) times per day, produced 32.25 ± 6.81 kg of milk per day, and consumed 5.03 ± 0.95 kg of concentrate in the AMS per day. The number of cows per robot box was 70.25 ± 14.06 and it was greater for guided flow compared with free flow AMS farms (75.0 vs. 65.5 cows per robot). Forty farms had exclusively free flow cow traffic. Milking speed was 2.24 ± 0.40 L/min and total milk yield per robot unit was 1843.5 ± 490.9 L/day. Total daily milk yield per robot has been suggested to be an important characteristic to assess the efficiency of AMS. Proc CORR in SAS was used to conduct a preliminary analysis of factors associated with yield per robot and milk yield per day. Factors most strongly associated (P < 0.001) with yield per robot were milk per cow per day (r = 0.81) and average milking speed (r = 0.76). Other factors moderately associated (P < 0.001) with yield per robot were average concentrate consumed per cow per day (r = 0.31) and exit length from the AMS (r = 0.32). Protected AMS exit lane was 3.1 ± 2.4 m long and ranged from 0.3 to 8.5 m. Factors associated (P < 0.001) with average daily milk yield per cow were milking speed (r = 0.79), exit length (r = 0.51) and average concentrate consumed per cow per day (r = 0.41). Further multilevel regression analysis will provide a clearer picture of factors influencing efficiency of AMS in the United States.