# Industry scenario matrices

## SaaS: tenant and entitlement boundaries
| Scenario | Expected rule to agree | Evidence |
| --- | --- | --- |
| Member reads another tenant's object | Deny without disclosure | Role/object matrix |
| Subscription changes an entitlement | Access changes at the specified boundary | State-transition results |
| Repeated billing callback | One agreed business effect | Integration evidence |
First engagement: a critical-journey and tenant-boundary review. Keep unrelated billing-provider certification outside its scope.

## Healthcare: traceability with synthetic records
Link each requirement to a test, build and reviewed result. Exercise clinician, administrator and patient-like test roles with fabricated data; verify refused access and record changes. Identify untested clinical decisions explicitly. First engagement: a traceability and access review of the named application layer, with specialist validation separately scoped.

## Finance: precision and repeat requests
Check integer minor-unit or agreed decimal precision, rounding boundaries, refused roles, expired authentication and repeated requests using isolated accounts. Compare the observable business effect as well as the HTTP response. First engagement: a focused contract and negative-scenario suite for the named API; keep claims about the financial organisation outside this test scope.

## E-commerce: checkout variations
Enumerate supported payment methods, countries, tax/delivery rules and viewports. Cover cancellation, provider refusal, retries, basket changes and keyboard navigation. Record the rules governing totals independently of the code. First engagement: a checkout release plan with performance and accessibility observations, not a guarantee of payment-provider availability.

## Logistics: partner and event state
Exercise malformed records, duplicate delivery, delayed and out-of-order events, retries and unknown identifiers. Record the sandbox differences from the production partner. First engagement: a partner-contract and event-state assessment with a controlled replay data set.

## Manufacturing: reconciliation and shop-floor clients
Compare source and destination counts, identifiers and units; report rejected and transformed records separately. Check the supported shop-floor browser/device combinations and interrupted connectivity. First engagement: an integration reconciliation worksheet and targeted web-interface checks; device firmware and safety validation require separate expertise.

## Startups: a bounded launch decision
Name the three journeys that carry the greatest user or commercial consequence. Confirm ownership, access boundaries and recovery. Separate launch-critical checks from improvements that can wait. First engagement: the readiness assessment followed by a small release evidence plan, with explicitly deferred automation or feature work.

For every sector record: build, expected rule, starting data, observed result, evidence, limitation, owner and review date.