Every decision is explainable. Every decision creates evidence.
A decision, a reason code, an outcome. Not a score, not a probability, not a confidence interval that someone has to interpret under pressure.
Three outcomes
The request satisfied the governing policy on every verified input.
A specific condition failed, and the reason code names which one.
The decision is held for a further condition or an authority to resolve.
One call. One answer. One record.
| decision | FAIL |
| rule | EBT Geofence (trafficking) |
| event | benefit-redemption |
| subject | demo:snap:trafficker-200 |
| when | 2026-07-31T12:31:50 |
| location | 39.46898, -99.28134 |
| reason_codes | GEO_OUTSIDE_ALLOWED_GEOFENCE |
The published list of reason codes with their meanings, so an integrator can handle each outcome without guessing at intent.
Evidence makes accountability possible
The record is not a log line written after the fact. It is the decision itself, retained with the inputs that produced it.
The record holds.
Decision history is retained so that what was decided cannot quietly change afterwards.
Every outcome is retrievable.
Any past decision can be produced with its rule, its inputs and its reason codes intact.
Attribution becomes a lookup.
Answering "who decided, and on what basis" stops being a reconstruction exercise.
The evidence surface: chain integrity and retrievable decision history.
Produce the record.
Bring a decision you have had to defend, and we will run it.
Request a demonstration