Skip to content

Product

Claims Document Intelligence built for the documents that sit between submission and payment

VerifyReceipt gives adjusters and claims leaders evidence they can act on, not a black-box score. The forensic layer between submission and payment.

Seven evidence layers in one workflowDuplicate, tamper, and claim-fit checksReviewer-ready next steps, not raw scores

What buyers want to know

Why this lands with claims teams

1

It makes the document legible before it makes it risky

Extraction, arithmetic, dates, duplicates, and image cues come back as a reviewer-usable case file.

2

It explains the issue clearly enough to act

The platform says what is wrong, why it matters, and what the reviewer should inspect next.

3

It fits real operations

Single uploads, API ingestion, and batch workflows all feed the same evidence-first review path.

Seven layers

One evidence stack, multiple claim-document workflows

Layer 1

Document forensics

PDF metadata, structural anomalies, authoring tool clues, and signs of flattening or export manipulation.

Layer 2

Image forensics

Compression artifacts, rendering inconsistencies, screenshot laundering, and region-level anomalies.

Layer 3

Structured extraction

Provider, dates, totals, currencies, and line items extracted with reviewer-safe confidence context.

Layer 4

Consistency checks

Arithmetic, tax, date, and claim-context validation to surface what does not add up cleanly.

Layer 5

Duplicate detection

Exact, near-exact, and reference-linked reuse signals across the insurer's own submissions.

Layer 6

Linked-case intelligence

Provider patterns, entity resolution, and repeated claim signals that help investigators connect the dots.

Layer 7

Human review workflow

Evidence bundles, reason codes, and operator actions that turn analysis into a documented decision.

Capability map

What VerifyReceipt actually does once a claim document lands

This is not just OCR. The platform reads the document, checks whether it fits the claim, flags reuse or manipulation, and turns the result into an explainable workflow your team can actually act on.

01

Read the document into structured evidence

Extract provider, dates, totals, currencies, line items, and document metadata so the reviewer starts with facts instead of manual transcription.

02

Check arithmetic, dates, and claim fit

Flag totals that do not add up, dates that do not make sense, and amounts or countries that do not fit the claim context cleanly.

03

Detect duplicate and repeat submissions

If the same PDF, a near-duplicate document, or the same invoice reference appears again, VerifyReceipt flags it and shows the earlier submission to compare.

04

Surface tamper and rendering anomalies

Use document and image forensics to highlight flattening cues, screenshot laundering, suspicious regions, and authoring inconsistencies.

05

Link related cases and entities

Connect repeated providers, entities, and case patterns across the insurer's own history so reviewers can see more than one isolated document.

06

Route uncertain cases into human review

Return plain-language reasons, what-to-check guidance, and a recommended next step instead of a black-box score alone.

07

Handle operational bulk workflows

Support dashboard upload, API-driven ingestion, and batch processing with per-file retries, progress tracking, and failure isolation.

08

Keep the decision trail auditable

Preserve reviewer actions, corrections, adjudications, webhooks, and retention-aware records so the result is explainable after the fact too.

Workflow

One operational flow from upload to documented outcome

1

Upload through API, dashboard, or batch ingest

Documents enter the same verification pipeline whether they arrive one at a time or in bulk.

2

Run the evidence stack

Forensics, extraction, duplicate checks, and claim-fit validation produce one evidence-led case file.

3

Highlight what matters

The portal surfaces the strongest issues first, including matched prior documents and what the reviewer should inspect.

4

Resolve in review when needed

Operators open the source file, compare evidence, apply corrections, refer, or adjudicate with a clear trail.

5

Report, callback, and audit

Results feed reporting, webhooks, and future rollout decisions without losing governance context.

What comes back

What the reviewer actually gets back

  • extracted fields and normalized claim facts
  • duplicate alerts with the matched prior document
  • plain-language reasons and what-to-check guidance
  • review routing and operator actions
  • audit-ready history for the final decision

Integration

Three ways to integrate

Production

Real-time API

Your claims system requests an upload URL, uploads the document directly to storage, then creates a verification with claim metadata. The workflow is async-first and designed for evidence-rich results.

Pilot + Production

Batch Retrospective

Upload a set of historical claims. Process overnight. Get an evidence-led report showing which claims were suspicious, duplicated, or worth re-examining.

Pilot

Dashboard Upload

Claims ops uploads a document directly. Sees verification status, evidence items, and timeline without a claims-system integration.

Wave 1 API sequence

POST /v1/documents/upload-url
X-API-Key: vr_live_...
Content-Type: application/json

{
  "filename": "receipt.pdf",
  "content_type": "application/pdf",
  "claim_ref": "CLM-1024"
}

PUT {upload_url}
Content-Type: application/pdf

POST /v1/verifications
X-API-Key: vr_live_...
Idempotency-Key: 7d613fe7-5243-4a6f-bc47-c7fe2d93b3a1

{
  "document_id": "doc_123",
  "claim_ref": "CLM-1024",
  "claim_type": "medical_invoice",
  "claimed_country": "TH",
  "claimed_amount": 15000,
  "claimed_currency": "THB"
}

Evidence-first means every decision stays explainable

Adjusters should be able to open the document, see what triggered the review, understand why it matters, and trace the technical basis only when they need it.