Current policy A
- Policy
routing-policy/v3- Historical selection
- Provider B
- Replay cohort
- ELIGIBLE
Payment routing change assurance
Replay routing changes before rollout. Verify what changed.
Compare current and proposed routing policies against the same replayable historical decision cohort. See which provider selections change and which decisions cannot be reproduced — without moving money or calling providers.
Historical replaySelection deltasExplicit replayability gaps
Know what a routing change will changebefore production does.
A proposed routing policy can be evaluated against the same historical decision state as the current policy. Nelvoss reports which provider selections change within the replayable cohort.
Decisions with incomplete policy, eligibility or decision-time context remain explicit replayability gaps, not inferred results.
Compare policy A with policy Bon the same replayable cohort.
Proposed policy B
- Same decision state
- Provider AProvider BProvider CProvider D
- Policy
routing-policy/v4-proposed- Proposed selection
- Provider C
- Selection result
- CHANGED
Cohort results are reported as CHANGED, UNCHANGED or UNREPLAYABLE — with an explicit missing-state reason.
The instruments that makea decision reviewable.
- COVERAGE
Replayable Historical Cohort
Separate decisions with sufficient decision-time state from those that cannot be reproduced.
- REPLAY
Policy Selection Replay
Evaluate current and proposed routing logic against the same supported historical decision state.
- CHANGE
Provider-selection Deltas
Report which selections change and which remain unchanged within the replayable cohort.
- GAPS
Verified Decision Coverage
Measure replayability and expose missing policy, eligibility or decision-time context explicitly.
- EVIDENCE
Evidence Bundles & Assurance Chains
Package verification results into reviewable artifacts whose altered or missing records are detectable.
- OUTCOME
Historical Outcome Analysis
Connect recorded selections with observed historical outcomes where the supplied evidence supports it.
Verification explainswhy the decision happened.Outcome analysis showswhat happened next.
Selection replay shows how routing logic changes. It does not establish the unobserved outcome another provider would have produced. Historical outcomes are shown only where recorded evidence supports them.
Demonstration / Synthetic example
Historical outcome comparison
Recorded decision
- Selected
- Provider B
- Policy
routing-policy/v3- Replay
- MATCH
Observed outcome
- Completion
- SUCCESS
- Expected cost
- 1.82%
- Observed cost
- 1.94%
Historical context
- Provider A
- 1.88%
- Provider B
- 1.94%
- Provider C
- 1.76%
Policy selection counterfactual: what policy B would select. Outcome counterfactual: whether provider B would actually approve or cost less. Ordinary historical observation does not establish the second.
AssuranceOutcome intelligence
NELVOSS
Pilot baseline
- Deterministic pilot acceptance
- PASS
- Verified Decision Coverage
- 100%*
- Decision Replay
- MATCH
- Evidence Bundle
- COMPLETE
- Real provider calls
- 0
- Transaction execution
- BLOCKED
- Orders
- BLOCKED
- Signing
- BLOCKED
- Payouts
- BLOCKED
A controlled way to find outwhat your routing records can actually prove.
Routing-change fit
Review the routing change, decision model and provider-selection surface.
Schema-first preflight
Start with a field dictionary or synthetic/redacted examples to assess structural replayability.
Controlled historical validation
If fit is confirmed, agree a bounded historical cohort or customer-hosted analysis path.
Replay & change report
Return replayability coverage, exclusions, historical replay and provider-selection deltas.
NO ORDERSNO PAYOUTSNO WALLETSNO SIGNINGNO TRANSACTION BROADCASTS
Discuss a routing changeQuestions that define the boundary.
No. Nelvoss operates as an assurance layer. Orders, wallet access, signing, payouts and transaction broadcasts remain outside the enabled pilot capability set.
No. The process begins with a field dictionary or synthetic/redacted examples. Provider credentials are not required.
No. Nelvoss verifies the integrity, completeness and reproducibility of the decision context it receives. Source-data truth remains a separate responsibility.
No. Selection replay shows what a proposed policy would select within the replayable cohort. It does not establish whether an unselected provider would actually have approved, cost less or produced a better outcome.
No. Nelvoss provides decision evidence and assurance tooling. It does not issue regulatory certification or compliance guarantees.
The published 100% figure applies to the deterministic Nelvoss pilot fixture. External production coverage can only be measured from the customer's actual decision records.
Routing change assurance / Fit discussion
What would your next routing changealter within the replayable cohort?
Start with the change, decision model and field dictionary. Historical data follows only after replayability fit and the data path are agreed.