What is OCR (Optical Character Recognition)?
OCR converts document images into machine‑readable text — IDs, bank statements, utility bills, paystubs. For identity flows, OCR accelerates capture and reduces typos; for review queues, it enables validation and comparisons. Quality varies with lighting, blur, fonts, holograms, camera sensors. Garbage in, garbage out.
Make it reliable: Prompt users to frame documents cleanly, detect glare, dewarp, and validate extracted fields with checksum rules (MRZ, barcodes) before accepting them. Compare OCR results against document templates and external sources, where permitted. For outcomes gating high‑value decisions, combine OCR with end‑to‑end identity verification so you’re not approving a crisp fake.
