Claim Flow AI public knowledge base

Claim Flow AI receipt intelligence field catalogue

This public field catalogue documents Claim Flow AI receipt and expense concepts. It contains no personal receipts, company reports, account records or customer statistics. Field extraction varies by receipt and needs human review.

Documented entities and relationships

This catalogue explains the data relationships used by Claim Flow AI. It contains no customer receipts or account records.

Receipt: Source documentLinks evidence to extracted merchant, date, amount, currency, and line-item fields.
Expense: Reviewable recordLinks a receipt or mileage record to a claim.
Claim: Submission boundaryGroups selected expenses into a finance-ready report.
Organisation: Optional destinationReceives a submitted report through an authorised company workspace.

Receipt-backed field groups

Merchant and venue identity
Transaction date and time when visible
Original amount and currency
Expense category
Line-item descriptions when legible
Location text when printed on the receipt
Dining and transport context only when receipt-backed
Review confidence and missing-field checks

Provenance and interpretation

Every extracted value should retain a relationship to its source receipt and review state. Aggregated statistics should publish their measurement definition, date range, population, and update time alongside the value.

Frequently asked questions

Does this dataset contain customer receipts?

No. It is a public documentation catalog of supported concepts and field groups.

Can field availability vary by receipt?

Yes. The receipt format, image quality and supported extraction affect available fields. AI can misread or infer values, so compare them with the original and correct mistakes.

How should statistics be published?

Each statistic should identify its metric definition, unit, population, time period, source, and last update.