When your Faisal Islamic Bank of Egypt statement arrives as a scan or a photo of the page rather than a clean export, the numbers on it are just pixels until they are extracted. Statements that arrive as a photograph or a scan, rather than a clean digital file, carry extra risk. The engine applies optical character recognition to lift the transaction table from the image. Once the numbers and text are read, the running balance is verified line by line against the statement's own printed closing figure. A single misread digit is enough to break the balance, and when that happens the row is flagged so you can correct it without reprocessing the entire document. The rest of the extraction stays intact.
If a transaction description on your Faisal Islamic Bank of Egypt statement is long enough to wrap across two or three lines, it still needs to land in a single row once converted. Long transaction descriptions that wrap across several lines on the page create a different problem: a single payment can look like two or three separate rows to a naive parser. The engine detects run-on lines and recombines them into one clean record, so the payment, date and amount stay together. The recombined row is still checked against the running balance, just like every other transaction, before the mapped columns are delivered in your chosen format.