OCR. The step that turns a scan into text for a program.
OCR (optical character recognition) is the technique that recognises the letters inside the image of a page, such as a scan or a photo, and turns them into text that can be searched, copied and read by another program. It is the first step for a paper document to become data.
OCR returns the text of the page. Which number is the total and which the customer code is decided by a later step, data extraction.
In Italy invoices between businesses do not need it: since 1 January 2019 they are electronic only and arrive with the data already structured. What remains are delivery notes, orders, documents from foreign suppliers, receipts.
Among Italian enterprises using artificial intelligence, 70.8% use it to extract information from text documents, according to Istat for 2025. It is the most common use.
This entry is part of the glossary of AI and automation, where every term has a short definition. Here the definition widens: how OCR works, the difference between a native and a scanned PDF, where an Italian business needs it, and how you get from text to data.
What is OCR
OCR is optical character recognition: a program looks at the image of a page, identifies the shapes of the letters and transcribes them as text. The result is a file in which you can search for a word, copy a sentence, or from which another program can read the data. Without OCR, a scan stays a photograph.
The technique has existed for decades, and in recent years artificial intelligence models have made it much more robust on crooked scans, phone photos, stamps and handwritten block capitals. Cursive handwriting, faded copies and tables with many columns and merged cells remain difficult.
Native PDF or scanned PDF
A PDF can contain two different things, and only one of the two needs OCR. A native PDF comes from a program, such as a document exported from management software: the text is already inside, and it can be selected and copied. A scanned PDF is the image of the page: to the eye it looks the same, but for a program it contains only coloured dots.
| Document | What it contains | Does it need OCR? |
|---|---|---|
| Electronic invoicebetween Italian businesses | Structured data, transmitted through the Sistema di Interscambio. |
No: the data is read directly. |
| Native PDFexported from a program | Real text, with the layout. |
No: the text is extracted as it is. |
| Scanned PDFfrom the scanner | The image of the page. |
Yes. |
| Phone photodelivery notes, receipts | The image, often crooked or in shadow. |
Yes, after straightening and cleaning the image. |
| Handwrittenforms, notes | The image of the handwriting. |
Yes, with variable results: block capitals hold up, cursive less so. |
A simple way to tell which kind of PDF you have: try to select a word. If only that word highlights, the text is there. If the whole page selects as a single block, you have an image, and reading it takes OCR.
Where an Italian business still needs it
Since 1 January 2019 invoices between parties resident or established in Italy «can only be electronic invoices», as the Italian Revenue Agency reports. In 2025, 2.4 billion of them passed through the Sistema di Interscambio, issued by 5.5 million parties. For these OCR is not needed: the data arrives already structured.
It is needed for everything else. Delivery notes and transport documents printed and signed. Orders arriving as a scan or a photo. Invoices from foreign suppliers, for which electronic format is optional. And then receipts, certificates, product sheets in PDF, contracts signed in pen and scanned again.
From OCR to data: extraction
A business needs to know which number is the total, which the date and which the item code: this second step is called data extraction. Today it is often done by language models, LLMs, which read the recognised text as a person would and fill in the fields for the management software.
It matters because retyping by hand has a small, constant error. In the Barchard and Pace study about 1% of data transcribed by hand contains an error, and checking by eye does not reduce it. Out of a thousand values retyped, that is about ten wrong numbers, within a plausible range, and therefore invisible.
How to judge an OCR system
An OCR and extraction system is judged by two numbers: how many fields it extracts correctly, and what it does when it cannot read. The first is measured on a batch of the company's real documents, counting the fields one by one. The second weighs more, because a total invented with confidence ends up in the management software and from there on the invoice.
The test to ask of whoever proposes a system to you is simple: give it an unreadable document and watch what happens. A well-built system sets it aside and flags it to a person. A badly built one fills in the fields anyway, and that is a hallucination in the format of data.
How Itria uses it
Itria's automatic document entry reads orders, delivery notes and invoices in PDF or as photos, with OCR where needed, and extracts the data in a format the management software accepts. On a test bench on 4 September 2026, with 30 documents and 237 fields, the document type was recognised in 30 cases out of 30 and 234 fields were correct, 98.7%.
In the batch there was an unreadable PDF, put there on purpose, with no text inside. The system classified it as «other» and did not invent a single field. It is a test on a test system, not at a client, and it can be rerun in front of whoever asks.
Related terms
ERP
The management software that links orders, stock and invoices. That is where the data extracted from documents ends up.
LLM
The language model that reads the recognised text and decides which number goes in which field.
Unstructured data
Free text, images, emails: information a program cannot use until someone puts it in order.
Data entry
Retyping by hand, which OCR and extraction are there to remove.
Questions and answers
What is OCR?
OCR, optical character recognition, is the technique by which a program looks at the image of a page, such as a scan or a photo, recognises the letters and transcribes them as text.
After OCR you can search for a word in the document, copy a sentence from it, or have another program read its data. Without OCR a scan is just a photograph.
What does a PDF with OCR mean?
It means a scanned PDF to which the text recognised by OCR has been added beneath the image. To the eye the page looks the same, but the words can be selected and searched.
To find out whether a PDF already has text, just try to select a word: if it highlights, the text is there; if the whole page selects as one block, it is an image.
Do electronic invoices need OCR?
Not Italian ones. Since 1 January 2019 invoices between parties resident or established in Italy are electronic only, as the Italian Revenue Agency reports, and they arrive with data already structured through the Sistema di Interscambio.
OCR is needed for delivery notes, orders as scans or photos, invoices from foreign suppliers, for which electronic format is optional, receipts and certificates.
How accurate is OCR?
It depends on the document: on a clean print it is very accurate, on faded scans, crooked photos and cursive handwriting much less so. The number that matters for a business is how many fields are extracted correctly on a batch of real documents, counted one by one, and what the system does when faced with an unreadable document.
As a benchmark, manual transcription gets about 1% of data wrong according to Barchard and Pace.
What is the difference between OCR and data extraction?
OCR turns the image into text. Data extraction reads that text and decides which number is the total, which the date, which the item code, and puts them in the right fields.
Today extraction is often done by language models, which read the text as a person would. A business needs both steps.
Notes on sources
- The quotation on the electronic invoicing obligation comes from the Italian Revenue Agency, read on 26 September 2026, which also specifies that for invoices to and from abroad electronic format is optional. The 2.4 billion invoices and the 5.5 million issuing parties are in the Agency's document I risultati 2025, pages 3 and 25.
- The 70.8% comes from Istat, Imprese e ICT, 2025, and is calculated on enterprises with at least 10 employees that use artificial intelligence.
- The error of around 1% comes from Barchard and Pace, Preventing human error: The impact of data entry methods on data accuracy and statistical results, Computers in Human Behavior 27(5), 2011: an experiment on people transcribing research data, cited for the shape of the error.
- The 98.7% and the unreadable document come from an Itria measurement on a test bench on 4 September 2026: 30 documents, 237 fields, the 3 wrong ones listed one by one.
We start from the documents someone retypes today. One real batch is enough to measure what they cost.
The first step with Itria is a fifteen-minute video call: we look at which documents arrive in the company every week and how many are still retyped by hand. Write us a line about what weighs on you. We take the first step: what a customer sees when they look for you, and what we found there. Even if we don't end up working together.