When to re-extract a document

Re-extract is designed for one core scenario: a rule changed after a document arrived, and you need the document to be processed under the new rule. It is not a general-purpose fix for every data problem.

Use Re-extract when

You added or changed a keyword rule. Documents that arrived before the rule exists aren’t affected by it. Re-extract them so the new routing applies.
You changed the first-approval rule for a supplier. Re-extract any documents from that supplier still in the flow.
A project was added to the ERP and synced to Nordflow. Documents captured before that project existed couldn’t match it; re-extract to give them another pass.
A custom capture rule was adjusted by the ECIT Digital team. Re-extract affected documents so the corrected logic applies.
A supplier card coding error was fixed. Re-extract existing documents from that supplier to pick up the new defaults.

Do not use Re-extract when

Situation	Better action
A single field value is wrong (supplier name, amount, date)	Edit the field directly on the document — faster, doesn’t reset the flow
The document has already been approved by several approvers	Edit in place or reject back to the service center — re-extracting forces everyone to re-approve
The source PDF itself changed	Reject back to the service center so they can decide whether to resubmit as a new document
The document just looks slightly off	Check if it needs fixing at all before resetting it

How re-extraction works in the background

When you click Re-extract, Nordflow removes the document from its current position in the approval flow and sends it back to the data capture engine. The engine re-reads the original PDF using all rules currently active for the client — custom fields, supplier whitelists, project patterns, first-approval routing. The document is then reinserted into the platform exactly as if it had just been received. Its approval history up to that point is preserved in the audit log, but its position in the flow resets to the beginning.