What Is Optical Character Recognition (OCR)? SharePoint Definition and Guide
Optical Character Recognition, or OCR, is a Microsoft 365 document processing service that extracts printed or handwritten text from supported images, scanned PDFs, hybrid PDFs, TIFFs, and embedded Office-file images. In SharePoint, the extracted text can improve search and make previously image-only content available to supported compliance controls.
What Optical Character Recognition (OCR) Means in SharePoint
When OCR is enabled, newly added eligible files are processed and their recognized text is indexed. For supported image files, SharePoint can expose the result in an Extracted text metadata column; for PDFs and TIFFs, text is indexed for search without necessarily appearing in that column. Microsoft documents extensive file, size, image-dimension, and language support.
OCR converts visual characters into machine-readable text; it does not necessarily understand document meaning or produce perfect transcription. Handwriting, skew, low contrast, stamps, complex tables, unusual fonts, damaged scans, and mixed languages can reduce quality. Structured extraction or human review is still needed when a field must be accurate.
Why It Matters to the Business
OCR can unlock archives, invoices, forms, drawings, labels, and scanned correspondence that employees cannot otherwise search. It also allows compliance tools to inspect recognized text. This can improve discovery while exposing sensitive information that was previously hidden inside an image, so search and security governance should be reviewed together.
Process owners should define whether the goal is findability, compliance inspection, full transcription, or downstream extraction. Searchable text may be sufficient for discovery but inadequate for payment, legal, or safety decisions. Measure recognition accuracy and user success on representative files rather than relying on a clean demonstration image.
What SharePoint Administrators Need to Know
Admins configure the service and pay-as-you-go relationship according to current Microsoft guidance. OCR processes newly uploaded content after enablement, so plan separately for existing archives. Confirm supported SharePoint site types, file extensions, dimensions, size, languages, Office embedded images, and de-duplication behavior.
Pilot with difficult real documents and test search, DLP, retention, sensitivity, and eDiscovery scenarios. Monitor volume and billing. Protect extracted text like the source file because it can contain personal or confidential data. Document when a human must verify recognition before a workflow uses the text.
What to Consider
Test difficult scans
Include handwriting, skew, low resolution, stamps, tables, mixed languages, faint text, and hybrid PDFs in the benchmark set.
Separate search from extraction
Text that is good enough to find a file may not be accurate enough to populate a consequential business field.
Plan existing content
Microsoft documents processing of newly uploaded files after enablement; define a controlled approach for legacy archives and their consumption cost.
Practical Takeaway
Use OCR to make image-based content searchable and inspectable, but validate accuracy before using recognized text in consequential automation or decisions.
How This Shows Up in the Field
Scanned contract archive
OCR makes clauses and party names discoverable in old PDFs, while legal staff still verify exact language against the image before relying on it.
Safety-form intake
Handwritten forms become searchable and route to a review queue; critical incident fields require a person to confirm the recognized text.
OCR and document processing models
| OCR | Recognizes printed or handwritten characters and makes the resulting text available for search and supported compliance scenarios. |
|---|---|
| Structured document processing | Uses a model to extract defined fields from documents with consistent layouts and writes values to structured outputs. |
| Unstructured document processing | Classifies variable documents and extracts information based on learned content patterns and examples. |
Related SharePoint Glossary Terms
Frequently Asked Questions
What files can SharePoint OCR process?
Microsoft documents scanned and hybrid PDFs, TIFFs, many image formats, and embedded images in supported Office files. File size, dimensions, site type, upload timing, and other limits apply.
Does SharePoint OCR process old files automatically?
Microsoft states that images uploaded after OCR is enabled are scanned. Plan and price a separate controlled approach if an existing archive must be reintroduced or otherwise processed.
Is OCR text always accurate?
No. Quality varies with scan resolution, handwriting, language, layout, contrast, damage, and typography. Use representative benchmarks and human review before consequential use.
Official Microsoft References
Editorial review date: July 13, 2026. Product status and licensing can change; confirm current Microsoft documentation before implementation.
Put Optical Character Recognition (OCR) to work with a clear plan
6SC can help your organization evaluate, design, implement, govern, and support Optical Character Recognition (OCR) in a maintainable Microsoft 365 environment.
Talk with 6SC