Finance Intelligence
Nothing here decides anything. Detectors produce scored observations with their evidence attached; a person opens the case and records what turned out to be true.
15 steps · 10 questions · 10 screens on this shelf
The screens on this shelf
- Risk Queue What is open, what it is worth, and how quickly it is being worked — sorted by exposure rather than by score,
- Risk Signals A scored observation with its EVIDENCE attached. "Score 0.86" persuades nobody; the facts that produced it are
- Risk Cases A person’s queue, with an outcome. The outcome is what closes the loop: confirmed fraud and dismissed noise ar
- Case Actions What was done about it in the world, as opposed to in the case.
- Risk Detectors Each answers one question, is separately switchable and separately measured. A detector nobody can turn off is
- Detector Performance Precision — of the signals it raised, how many were real. The only number that decides whether a detector shou
- Assistant Suggestions Suggestion rather than detection, with the reasoning attached. "Because the last nine invoices from this suppl
- Coding Patterns A learned lookup rather than a model. For GL coding this beats a model on both accuracy and explicability, and
- Narratives The generated text and the edited one, both kept — the difference between them is the most useful feedback the
- Assistant Accuracy Confidence is only useful if it is calibrated: a suggestion offered at 0.9 and accepted half the time is worse
Progress is kept in this browser only — it is not sent anywhere and nobody else can see it.