Automation and integration
OCR (optical character recognition)
OCR (optical character recognition) is technology that converts images of text, such as scans, photos and PDFs without a text layer, into machine-readable text.
How it works in practice
OCR finds the characters on a page and returns them as text, often with their position. It is the first step in reading any document that is not already digital text, and its accuracy depends on scan quality, handwriting, fonts and language.
On its own, OCR gives you text and nothing about which number is the invoice total. That is the job of document processing models built on top of it. Many current vision-language models read the image and understand it in one step, but a dedicated OCR pass is still useful for searchable archives and for checking what the model saw.
A worked example
A clinic digitises paper referral letters. OCR makes every letter searchable in the patient record, and a document model pulls out the referring doctor and reason for referral so the booking team can triage them.
Examples in this glossary describe typical builds for companies of 1 to 1,000 people. They are not named client results.















