**Availability: blueprint example.** This article describes a possible workflow. It does not provide a live upload tool or promise that a specific document format is supported.

## The job

Turn scanned pages or document files into text and a small set of fields that a person can review.

| Part | Example |
| --- | --- |
| Input | A synthetic scanned form or document file. |
| Output | Extracted text, selected fields and a link back to the source. |
| Human review | Check uncertain text, field values and missing pages before accepting the result. |
| Possible next step | Add the reviewed record to an approved system or make the text searchable. |

## Where it can help

The useful part is reducing repetitive copying and making the source easier to work with. OCR reads visible text; it does not by itself decide whether a document is correct, complete or suitable for a particular decision.

## Questions to settle

Which file types and languages matter? Are pages handwritten? Do you need tables or only a few fields? How long should the source be retained? Who can open it? How will duplicates and failed jobs be handled?

Ask for a preview with a synthetic sample before treating the workflow as available for your own material.

Follow the [intake walkthrough](/hc/productpat/articles/productpat-en-intake-walkthrough) or bring this building block into your [solution brief](/hc/productpat/articles/productpat-en-solution-brief).