Knowella

The Flagged queue — 32 flagged, five reasons

Why a well-read document still gets flagged, the five flag reasons, My Carriers, the review screen, and clearing a flag.

3 min read Real product screens
app.knowella.com Live walkthrough
Flagged is where a document goes when KnowDoc read it well but couldn't finish the job: there's no load to match, a POD is missing, or someone on your team set it aside. The walkthrough queue holds 32 documents, every one extracted at 85% confidence or better. The reading worked. What's missing is a load, a POD or a page, and the review screen puts the fix next to the question.
In this article

Read the reason column before the count

The header says "32 documents need attention", but the Flag Reasons column tells you what kind of work that is. In the walkthrough, 32 flags come down to five reasons, and most of the queue is one of them. Check the Confidence column too: when every row is high, the problem isn't the extraction.

  • No load found — Most of the queue. Broker and factoring invoices (Ham Brokerage, for example) with no load-shaped reference to match on.
  • No load number — The invoice was read, but the load number field came back empty.
  • Missing POD — No proof of delivery attached. Two Arrive Logistics invoices in the walkthrough.
  • Incomplete POD — A POD is there but missing pages or a signature. One Uber Freight statement.
  • Human flagged — Someone on your team set it aside for a second look.

My Carriers — your slice of the queue

Turn on My Carriers to see only the carriers you look after. In the walkthrough the list drops to zero and shows "All caught up": nothing in this queue is waiting on this person. Turn it off to see the whole team's queue again.

Open the document

Open the ⋯ menu on a row. View gives a quick look, Detailed View opens the full review screen, and Download saves the original file.

The review screen

Detailed View shows the email the document came in (sender, subject, a Reply button) above three panes. Compare the AI Recommendation with Current Data, then check both against the original PDF.

  • AI Recommendation — What KnowDoc read from the invoice, and the flag it raised.
  • Current Data — What's already in the system for this load.
  • The PDF — The original document, with paging and zoom. It settles any disagreement.

Check related paperwork and why it was flagged

If other files came in the same email, they're listed under Related paperwork. In the walkthrough, 474943.pdf arrived with INVOICEi5673.pdf, so tick it and click Add to this freight bill. The red pill names the flag ("Flagged: generic_needs_load") and the Load Number field explains it: no load-shaped reference in the extraction, so there's nothing to look up. Further down, confirm there's no existing freight bill on the load and check the Load Documents list.

Clear the flag

When the document answers the flag, click Clear Flag. The flagged count and the sidebar badge both drop by one (32 → 31). Only clear a flag you've actually resolved; the next person will assume it was checked. For a Missing or Incomplete POD, reply to the carrier and ask for the paperwork instead.

Let the AI Agent summarise the queue

The AI Agent reads the whole queue for you: nothing is a reading failure, most flags are No load found, and three need paperwork from the carrier. It offers to draft the three POD requests (two to Arrive Logistics, one to Uber Freight) so you can send them in one go.

Run your messiest freight bills through it.

Send us a few of your hardest multi-page freight bills, and see the split, the checks and the audit for yourself.