Text that only looks readable
Some Arabic PDFs carry a text layer that copies out as nonsense because of legacy fonts. Each page's text is checked before it is trusted, and read again from the page image when it fails.
- Madar
- Document Intelligence
Scans in. Organised documents out.
Document Intelligence reads Arabic and English documents, separates combined files, classifies each document and prepares both the document and its data for your archive and the workflows that use it.
Arabic and English. Built in Kuwait.
One scan, four documentsSample
Scanned file
scan_0412.pdf · 12 pages, Arabic and English
Documents found
- Policy schedulePages 1–4
- Police reportPages 5–7
- Repair estimatePages 8–11
- InvoicePage 12
Ready
Four files, named by type and date, with their key fields indexed for the archive
How it works
Four steps from a scan to a filed document.
Reads
Printed Arabic and English text is recognised, including pages that mix both.
Separates
A combined scan is split into the individual documents inside it.
Classifies
Each document is labelled by type: a policy, an ID, a report, an invoice.
Prepares
The document and its data are named, indexed and passed to your archive or the next step in the workflow.
Arabic documents
Where Arabic OCR usually goes wrong, and what we check.
Two languages on one page
Bilingual pages are read block by block, so an Arabic column and its English translation both come out intact.
Mistakes that repeat
A misreading that repeats through a document can pass for a real word. Checks look for known error patterns, not only for unusual words.
Questions
Document archiving and Arabic OCR, answered.
Does it replace our archiving system?
No. It prepares documents and their data for the archive or document management system you already use.
Which languages does it read?
Arabic and English, including bilingual documents.
Can it handle a stack of documents scanned as one file?
Yes. That is what separation is for: one scanned batch is split into the documents inside it, and each is classified on its own.
Where do the documents go afterwards?
Into your archive and the workflows that use them, for example a claim case opened by Intimation & Workflow Automation.
Show us the documents that slow your team down.
What arrives, in which languages, and where it needs to end up. We will show you how Document Intelligence would read, split and file it.