Computer-Vision Insurance Card & ID Scanner for a Pediatric Practice

Overview

Built a computer-vision tool that lets parents photograph their child's insurance card and ID during online check-in, automatically extracting and pre-filling the relevant fields instead of requiring manual entry at the front desk. The system validates that key fields like member ID and group number were captured clearly and prompts a retake if the image quality isn't sufficient for reliable extraction. This was built specifically to fit into the pediatric practice's existing online check-in flow rather than requiring a separate app or portal. Extracted data is reviewed by front-desk staff before being saved to the patient record, keeping a human checkpoint in place. Front-desk staff at a pediatric practice were manually entering insurance and ID information from every check-in, a repetitive task that added friction to hundreds of weekly visits. We built a computer-vision tool integrated into the practice's existing online check-in flow that lets parents photograph their child's insurance card and ID, automatically extracting and pre-filling the relevant fields with a human review step before saving. We tested the extraction accuracy against a wide sample of real insurance card formats before integrating it into the check-in flow, tuning the retake-prompt threshold to balance accuracy against parent friction. The tool launched alongside the existing check-in process with front-desk review built in from day one, not added after the fact. Manual insurance data entry time dropped by an estimated 3 minutes per visit, a meaningful savings across hundreds of weekly pediatric visits, while a front-desk review step kept accuracy in check.