AI CXO
Value created · QTD₹4.82 Cr▲9:14:00 AM · Mon 26 Jan ISTActive

Scoring models

The engine reads these for you. Open one only if you want to see — or tune — how a score is built.

◍

This is the MRI. Every score in the DSOE Scoring & Trigger spec is here, with its weights, thresholds and SAP sources — fully inspectable and tunable. Day to day you never open this: the platform reads these signals and hands you the summary and the recommendation under Today, Executions and Dealer 360.

+ 5 more in spec

Dealer Health Score

health_score · 0–100 · high = healthy

Is this dealer in a stable, well-functioning trading relationship right now?

How to read this. Each slider is the weight — the relative say each signal has in the final score. The score is Σ (weight × sub-score), so raising one weight makes that signal matter more and the others less. Weights must total 1.00. Drag any slider and watch the sample dealer below recompute in real time — that's exactly what the engine would do to every dealer tonight.
Sub-score weights · SAP source
Order recency
0.22
SAP SD · VBAK-ERDAT
Order-frequency stability
0.20
SAP SD · COUNT(VBAK)
Payment discipline
0.25
SAP FI-AR · BSAD
Engagement (linked)
0.18
SaarthiOS · engagement_score
Complaint intensity (inv.)
0.15
SAP QM · QMEL
Σ weights = 1.00must total 1.00
Live preview · DLR-1002
13.6/100
At-risk
DLR-1002 = 0.22×0 + 0.20×0 + 0.25×40 + 0.18×20 + 0.15×0 = 13.6. Change a weight and this recomputes — and if it crosses a band threshold, the dealer moves in or out of the trigger.
Band thresholds
Healthy ≥
Watch ≥
At-risk <
◷ Recomputes nightly and on every relevant SAP event. Data-quality gate holds a score if a feed drops below its freshness SLA.
Demo — fictional companies, illustrative data