Intimate partner violence (IPV) accounts for a large proportion of police calls for service and criminal offenses in most jurisdictions and families often experience significant harm from these incidents. In an effort to rationally allocate investigative resources, police departments are turning to risk assessment scales to identify suspects and victims at greatest risk for further incidents. The existing measures suffer from a number of issues including limited predictive validity, lengthy scoring procedures, and the requirement of supplemental data collection. The current study uses a sample of IPV cases from a large US city to assess whether recidivism can be predicted using factors that are already available in most law enforcement records management systems (RMS). A series of multivariate regression analyses demonstrate that domestic recidivism was predicted (AUC = .72) using a combination of suspect, victim, and index offense characteristics obtained from the local RMS. Additional analyses found that victim-reported risk factors obtained via supplemental interviews did not further benefit predictive validity. The findings suggest that additional data collection for the purpose of IPV risk assessment may be unnecessary for most law enforcement agencies. Moreover, the use of currently available records supports the automation of risk triage tools, which could yield further improvements in predictive accuracy and decision-making efficiency.