In response to a request for information from the Office of Management and Budget (OMB) on the responsible procurement of artificial intelligence (AI) systems, Urban Institute experts recommend that agencies (1) require vendors to share detailed information on the representativeness, collection practices, and gaps in the data used to create their models; (2) require vendors to document their model selection and tuning processes, commit to continuous evaluation and reporting of AI systems, and agree to update processes as best practices evolve; (3) create concrete metrics for measuring equitable AI outcomes, prioritizing procedural, distributional, and structural equity; and (4) commit to soliciting more meaningful community participation and feedback.