Transaction Anomaly Detection
HealthyReal-time production health
Known immediately — no outcomes needed.
Predictions / hour
Latency
Inputs, context & predictions
Input drift vs reference (PSI)
Prediction distribution
Routing
- Backend
- Managed by Zero Grad
- Context
- pinned to v2
- Fallback on timeout
- Managed C · 0.02% of requests
Matured outcome quality
Performance metrics are computed only for predictions with observed outcomes. Analyst-confirmed incidents take up to ~2 days to arrive, so recent predictions aren't scored yet.
Business impact
Sep 1 – Sep 30 · 89% coverageExposure is the total value of events in confirmed incidents at time of alert.
Model quality
Sep 1 – Sep 30Alert precision — matured outcomes
By prediction day. Recent cohorts are provisional until their outcomes arrive.
Outcomes
Ground truth flows back by prediction_id. It improves evaluation and routing — never model weights.
Send outcomes
Every prediction returns a prediction_id. Send the real outcome whenever you learn it.
# Later, when the real outcome is known
zg.capability("ledger-anomalies").send_outcomes([
{"prediction_id": "pred_01J9X4T5WE", "outcome": {"incident": True},
"observed_at": "2026-09-30T23:41:00Z"},
])Outcomes never update model weights — Zero Grad does not train a customer-specific model. They are used to measure real production quality, compare foundation models and context-selection strategies, improve routing, and decide whether a challenger beats the current production configuration.
What Zero Grad learned from your outcomes
No training doesn't mean no learning.
- Analyst feedback: 38% of weekend payout alerts were false positives → a weekend baseline was added to context.
- TabFM with labeled incidents in context raises alert precision from 84% to 88% (v3 challenger).
- Both missed incidents were slow drifts → foundation-model residuals are now weighted higher.
What adapts: which foundation model serves which slice, how in-context examples are selected, and calibration. What doesn't: model weights.