Many water utilities have invested time and resources to establish an initial service line inventory per the Lead and Copper Rule Revision (LCRR), but many are faced with a daunting number of service lines with unknown material. Submitting unknowns does not violate the LCRR but will trigger additional work by the utilities and present tremendous challenges. Statistical Analysis and Predictive Modeling (also known as Machine learning) can be cost-effective tools to reduce the number of unknowns in the service line inventory and streamline the removal of lead service lines from our communities. Predictive Modeling, also known as Machine Learning, is an investigative method listed in EPA's LCRR Inventory Guidance (August 2022). Predictive Modeling and Statistical Analysis are accepted investigative methods by Kentucky Division of Water (KDOW). Tennessee Department of Environment and Conservation (TNDEC) accepts the use of Predictive Modeling. For systems that do not expect to have any lead or galvanized service lines, KDOW allows the use of a Statistical Analysis to demonstrate non-lead and significantly reduce the field verifications required. For systems that expect to have some lead or galvanized service lines, both KDOW and TNDEC allow the use of Predictive Modeling to reduce the unknowns which also significantly reduces the field verifications required. In this presentation, Trinnex will share case studies highlighting both applications and discuss the process of both methods. We will also share a utility’s Predictive Modeling result that achieved an accuracy of 90%. The audience will walk away with a new understanding of the tools available for tackling LCRR unknowns.