A B2B commerce and financing business is deciding where to draw the boundary between its commerce and credit domains. What should determine it?

What should determine it

• Different invariants. Commerce enforces "this order is valid and fulfillable"; credit enforces "this exposure is within limits and correctly recorded". Those are different truths with different correctness requirements, and mixing them means the strictest applies to both.
• Different rates of change. Catalogue, pricing and fulfilment change constantly; credit policy changes slowly and under regulatory constraint. Joint deployment forces the fast domain to move at the slow one's pace and the slow one to accept the fast one's risk.
• Different regulatory scope. Credit is regulated; commerce largely is not. Keeping the regulated perimeter small is a substantial simplification — every system inside it inherits audit, retention, reporting and change-control obligations.
• Different failure consequences. A commerce outage loses orders; a credit error loses money and creates regulatory exposure.

The integration between them

Explicit and asynchronous where possible. Commerce publishes events — order placed, delivered, returned; credit consumes them and maintains its own view of exposure.

Credit must own its own record of the facts it depends on, rather than querying commerce at decision time, because a credit decision must be reproducible and auditable years later and a live query against a system that has since changed is neither.

The tension to resolve deliberately

The customer experiences one product. A buyer wants to place an order and have credit applied in one flow, and the boundary must not make that clumsy.

The resolution is a synchronous read of a pre-computed credit position — available limit, maintained by the credit domain — with the authoritative decision and its record staying inside credit. The commerce flow gets a fast answer; the regulated record stays where it belongs.

The failure to avoid

Credit logic leaking into commerce for convenience — an available-limit calculation duplicated in the checkout flow because it was faster than an integration. That duplication diverges, and the divergence is discovered as an exposure discrepancy in a regulatory report, which is the most expensive possible way to find it.