Scanning a Document to Create One

Norma can take a document you already have on paper or as a file — a supplier invoice, a purchase order, a delivery note, a photo — and turn it into the matching SAP document, without you retyping any of it. This is called intake (UC-07). This page explains the flow, what you review, and which document types are supported.

The flow, step by step

1. Upload. You attach or upload the document. Norma stores the raw file safely under your identity; the file itself never travels through the rest of the pipeline — only a reference to it does, so large scans stay out of the workflow history.
2. Extract. Norma reads the document — text, layout, and, for a spoken instruction, the audio — using its extraction service (docling). Out comes the raw text and structure.
3. Structure. Norma works out what kind of document this is (invoice, purchase order, delivery, …) and maps what it read onto the fields that document needs: the business partner, each line item and quantity, and the amounts.
4. Review. The structured result becomes a proposed SAP document in the same review gate as any other write. You see the extracted fields laid out and editable.
5. Create. After you approve — correcting any field the scan misread — Norma creates the SAP document.

A scan is just another way to start a request. Once it’s structured, it flows through exactly the same reason-review-create path as a typed instruction. There is no separate, less-supervised route for scanned documents.

What you review

Because a scan can misread a digit or a name, the review step matters. You’ll see:

• the document type Norma detected,
• the business partner it matched,
• each line — item, quantity, price,
• the totals it extracted.

Every field is editable. Fix anything the scan got wrong, then approve. If a field is corrected, the created document and its audit record use your corrected values. If the extraction fails outright — an unreadable scan, unclear audio — the task ends as an error you can see, never a silent empty result.

Supported document types

Intake can classify and create the SAP documents Norma already knows how to write. Today that means:

• Sales order — the original intake type.
• Purchase order.
• Purchase quotation.
• Goods receipt PO.
• Delivery.
• A/R invoice.
• Quotation.
• Return.
• Credit memo.
• A/R down payment.

As Norma’s write set grows along the SAP document chain, intake recognises each new type the same way. Document types Norma cannot yet write (for example purchase invoices, purchase returns, or inventory movements) are not yet available from a scan either.

What intake does not do

• It does not skip the review gate — a scanned document is approved by a human like any other write.
• It does not create a document type Norma can’t otherwise write.
• It does not reason further about the data on its own; extracting and structuring the scan is where intake stops, and the normal pipeline takes over from there.

In short

Hand Norma a document, it reads and structures it, you check and fix the details, and it creates the matching SAP document — all through the one review gate that guards every write.