Open Standards
The schemas we publish, so you don’t need us to verify the numbers
A finance system is only as trustworthy as the formats it emits. We publish the data dictionary, ledger schemas and decision records ADmetric AI runs on — versioned, parseable, and signed.
Published standards
| ID | Standard |
|---|---|
| AT-001 | Pacing event schema Append-only event format for every plan, adjustment and platform sync. |
| AT-002 | Reconciliation ledger Double-entry ledger format mapping platform invoices to GL postings. |
| AT-003 | FX leak record Per-transaction FX corridor with mid-market reference and slippage. |
| AT-004 | Material decision record Structured artefact for every automated decision above materiality. |
| AT-005 | Control evidence pack Auditor-portable bundle for SOC 2 and ISO 27001 control testing. |
| AT-006 | Treasury policy file Machine-readable IPS for allocation bands and counterparty ceilings. |
| AT-007 | Platform fee taxonomy Normalised fee codes across Meta, Google, TikTok and Spotify. |
Read the spec
Every schema has a written spec with examples and edge cases.
Download the bundle
JSON Schema + Parquet + reference parser, signed by our root key.
Track changes
Public changelog and a 12-month overlap window for major versions.
Commitments
- Every standard is published under CC BY 4.0 with a public changelog.
- Breaking changes ship behind a major version with a 12-month overlap.
- A reference parser exists for every format we emit — in TypeScript and Python.
- Customer exports validate against the same schema we use internally.
Why we do this
Lock-in is a tax on customers and a moat for vendors who can’t win on merit. We’d rather lose a renewal than win one we couldn’t defend with open formats. See interoperability and our customer promise on portability.