Platform · Ontology
Financial ontology
One model instead of fifteen exports. People, organizations, accounts, transactions, documents, devices, employees, branches and alerts — connected, with provenance on every value.
01
Objects and relationships
Twelve relationship types, each carrying its own source references, confidence and validity period.
- OWNS
- CONTROLS
- REPRESENTS
- USES
- HAS_DOCUMENT
- SENT
- RECEIVED
- WORKS_AT
- SERVED_BY
- CONCERNS
- RELATED_TO
- PARTICIPATES_IN
02
Entity resolution proposes; a person decides
Records merge automatically only on deterministic identifiers such as a national id or a tax id. A shared phone number, a matching date of birth, a name spelled differently in two systems — these become proposals for review, never silent merges. Accepting one is reversible.
Merge proposal · example
D. Okafor (Core Banking) · Daniel Okafor (CRM) — shared phone number.
Proposed · awaiting reviewReversible
03
Provenance is not optional
Every field shows its source, the source system's own identifier, when it was extracted, when it was last updated, and a confidence signal. An analyst who cannot answer where a number came from cannot defend a conclusion built on it.
- Value
- 48,200.00 USD
- Source
- Core Banking · txn_export_v2
- Source identifier
- CB-TX-2026-088412
- Extracted
- 2026-08-08 21:52 UTC
- Confidence
- Deterministic
See it on your data’s shape — not a slide deck.
A pilot runs on synthetic or masked data first, inside your perimeter.