Background: Various organizations, including Centers for Medicare & Medicaid Services, have defined accountability measures that are publicly reported and used to adjust payment for clinical services. We posit that if primary care clinician (PCC) performance on accountability measures (such as diabetes or cardiovascular care) is suboptimal, then PCC-specific care improvement opportunities can be identified through automated monitoring of a broad set of specific underlying process-of-care measures pertaining to clinical actions. Methods: As part of a clinical decision support (CDS) system implemented within a large health care delivery system, we provided individualized patient treatment recommendations related to optimal management of 6 cardiovascular risk factors at adult encounters. We developed process-of-care measures related to recognition of conditions, timely monitoring of disease states, time to medication intensification for patients in suboptimal control, medication choices, frequency of drug/condition and drug interaction safety concerns, and appropriate referrals by calculating the provider’s proportion of patients that each of 60 clinical recommendations were generated by the CDS system (10 blood pressure, 15 glucose, 29 lipid, 4 aspirin, and 2 smoking). We then quantified variation across 597 PCCs (minimum 20 patients) on these process-of-care measures by analyzing the data from 128,679 clinic visits made by 75,855 adults with diabetes or cardiovascular risk factors from January 1, 2017, through December 31, 2017. Providers were classified as high-performing (10th percentile) and low-performing (90th percentile) for each recommendation based on the proportion of patients for whom the recommendation was generated. Results: We observed substantial variation when comparing low- vs high-performing providers, with as much as a 29% difference in the proportions of patients for whom CDS recommendations were generated. Across measures, low-performing providers had a proportion of patients for a CDS recommendation at least 1.9 times greater than high-performing providers. Conclusion: In a care delivery system frequently recognized for high-quality care, we observed many clinician-level opportunities to improve underlying processes of care that are linked to optimal diabetes and cardiovascular risk factor management. Wide variations observed in performance on process of care presents actionable opportunity to design and deliver PCC-specific quality improvement interventions. The dissemination potential for this automated surveillance technology is high.