Based on our Aug 25 discussion: the case is the provider (e.g. DHD).
Billing + claim documents feed document patterns, fee checks, and opinion-letter flags.
Pick a flow below and walk the steps.
How to use this page.
Click Flow A or Flow B, then use Next / Back through the steps.
Yellow/blue note boxes explain why each step exists. Nothing here saves to the real system.
For the proposed Drive + JSON pilot (sample data), see /ui/mock2.
For live data, see /ui/mock4.
1 · Select provider from the global book
You already know the target (e.g. DHD). Claimants are vehicles — the case is the provider.
Note. The large billing ingest stays for fee / volume analytics across many providers.
Deep document work starts by picking one provider, not by pasting docs for all 4,000.
4,280Billing lines in history
186Distinct claims
HighPrior risk signal (demo)
2 · Claims from this provider’s billing
Billing defines the claim list. Next you paste title + text onto claims, not every line.
Dummy data — three sample claims for DHD. In production this list comes from the ingested billing book.
Claim #
Accident
Lines
Billed
Docs
1000348
25-Jun-2017
24
$8,420
0 attached
1000991
03-Aug-2017
18
$6,110
0 attached
1001204
12-Sep-2017
31
$12,905
0 attached
3 · Paste documents onto each claim
After you pick the provider, billing already listed the claims.
For each claim you care about, enter a document title and
paste the text (from your OCR tool or claim file). No in-app OCR yet.
What we plan for now.
Investigator opens a claim → chooses type (MV-104, initial report, MRI…) →
types a title → pastes content → saves. Repeat for more docs on that claim.
Later: bulk PDF upload + OCR that auto-fills this same form.
Claim 1000348· 24 billing lines
Add a document to this claim
Claim 1000991· 18 billing lines
Add a document to this claim
Claim 1001204· 31 billing lines
Add a document to this claim
Why paste. Your team already extracts text (or will use an external OCR).
We store title + text on the claim so analysis can compare wording across DHD claims.
PDF upload + OCR inside the product comes later.
4 · Run analysis on this provider
One run covers document patterns, fee-schedule checks, and opinion-letter relevance for the services billed.
Document patterns across claims (boilerplate / protocol / intervals)
Billing vs fee schedule (overbilling, missing schedule, timeline)
Opinion letters matched to service type (e.g. PT letters for PT billing)
Demo. Click Run — we show canned results on the next step (no live AI).
5 · Results (dummy report)
What investigators take away: flags, story, exportable CSV, and later open/export source docs.