No date on the page? It doesn't vanish. It gets its own bucket.
Undated Document Flagging catches every document where no date can be extracted — a note with no header, a scan with no visible date stamp — and routes it to its own dedicated undated bucket, instead of guessing a date to force it onto the timeline or letting it vanish from the chronology. It stays visible until a date is confirmed.
Nothing disappears for lack of a date.
Every page in the upload gets a date-extraction pass. When a date can't be pulled from the header, the body, or a stamp, the document doesn't get skipped and it doesn't get forced onto the chronology at the wrong point — it's routed straight into its own undated bucket, tracked like every other document in the case.
A queue built for confirmation, not guesswork.
The undated bucket is a working queue, not a dead end. It stays a separate, filterable view until a reviewer confirms a date — so the chronology only ever shows dates a person stood behind, and the review workflow is exactly what it should be: confirm, don't hunt.
A silent drop and a false date cost the same thing: trust.
A chronology built on a chart with a guessed date is one cross-examination away from a problem. Medrecords AI never fills that gap silently — the undated bucket keeps the record honest about what it actually knows, and puts the confirmation decision where it belongs: with the reviewer.
See Medical ChronologyFrom an undated page to a confirmed entry.
4 steps, and not one of them involves guessing a date.
Undated pages arrive mixed in with everything else. No pre-sorting, no separate pile.
Headers, stamps, and body text are checked for an extractable date on every single page.
No date found means no guessing and no drop — the page lands in its own queue instead.
Once a date is confirmed, the document takes its place on the chronology automatically.
Who relies on the undated bucket.
Different reasons to care, same problem: a missing date shouldn't mean a missing document.
A missing treatment date can't quietly disappear from a demand package.
For PI firmsFaxed forms and clinic notes without headers get flagged, not lost, at volume.
For workers compA clean, filterable queue of exactly what needs a date confirmed — nothing more.
For LNCsNo opposing expert can point to a guessed date buried in the chronology.
For IME orgsRelated features.
The timeline every confirmed date lands on, synced to its source page.
See the chronologyA likely date of injury, suggested with citations — confirm instead of hunting.
See the featureFlags visits and providers that should exist but weren't produced at all.
See the featureA confirmed date re-sorts the case automatically — no re-run needed.
See the featureUndated Document Flagging, answered.
It's routed into a dedicated undated bucket instead of being assigned a guessed date or dropped from the case. It stays visible in the working file, flagged for review, alongside everything else that was uploaded.
No. Undated documents are never assigned a guessed date to force them onto the timeline — they stay flagged in the undated bucket until a date is confirmed by a reviewer.
In their own bucket, separate from the dated chronology, so they're easy to find and work through instead of being buried in the main timeline or missed entirely.
Yes. The undated bucket is its own filterable view, so a reviewer can work through exactly the documents missing a date without paging through the rest of the file.
It moves out of the undated bucket and takes its place on the chronology and index, same as any other dated document — no re-running the case.
Key Date Auto-Suggestion proposes a likely date for a specific case-critical date, like date of injury, with citations to confirm. Undated Document Flagging is the safety net for a page that carries no extractable date at all, so it's never silently lost.
Anything where a date wasn't captured on the page or in the scan — most often handwritten notes without a visible date stamp, cropped pages, or low-quality scans where the date field didn't OCR cleanly.
See what your undated bucket looks like.
Book a quick demo, or upload a single file and see exactly which pages get flagged instead of guessed. No commitment. Handled under our BAA; never used to train a model.