Measures confidence independently from the recommendation itself. Seven inputs — Data Quality, Data Freshness, Source Reliability, Historical Accuracy, Completeness, Conflicting Evidence, and Model Agreement — feed deterministic scoring across four confidence layers (Overall, Evidence, Prediction, Recommendation) and five bands from Very High to Insufficient Evidence. AIREx explains why confidence changed and recommends concrete ways to raise it.
Overall
67%
Evidence
80%
Prediction
75%
Recommendation
71%
Preset
Load a scenario
Subject
What is being measured
Signals
Adjust the seven inputs (0–100)
Data Quality82
Data Freshness78
Source Reliability85
Historical Accuracy74
Completeness70
Conflicting Evidence(penalty)35
Model Agreement76
Live deterministic preview: 67% overall · band Moderate
Confidence readout
Acquisition confidence — 1420 N Loop MOB at $184M
Moderate
Overall Confidence67%
Evidence Confidence80%
Prediction Confidence75%
Recommendation Confidence71%
AI narrative
Run the engine for full explanation
The gauge above uses the deterministic scoring formula in real time. Press Measure confidence to generate the change explanation, positive and negative drivers, and prioritized improvement actions with expected lift.