Scoring
Rule components, AI assessment, and the ignore / qualify / opportunity thresholds.
Each organization has an active scoring config (versioned) with three thresholds and a set of scoring rules. A signal gets a total from 0–100. Every component is stored with a reason so a person can see why the number moved.
Thresholds
| Threshold | Meaning |
|---|---|
| Ignore below | Score under this → status ignored; no resolution or enrichment |
| Qualify at | Score at or above this continues the pipeline |
| Opportunity at | Score at or above this may create an opportunity at the ready stage |
Home coordinates on the config are used by geographic-distance rules.
Dimensions
Rules and AI assessment write components on these dimensions:
- Commercial relevance
- Project size
- Expected duration
- Product fit
- Timing
- Geographic distance
- Contactability
- Confidence
Deterministic rules are parameterised in scoring_rules (rule_key, weight,
max points, JSON params). AI assessment typically contributes
commercial_relevance and product_fit. The total is clamped 0–100.
Industry profile
Personas (title keywords, priority) and products (fit keywords) on the organization industry profile feed both classification and enrichment. EventDock seeds construction/events personas and products such as camera security, WiFi, and access control.
Re-score from the signal when you change the active config or rules. Previous
score rows remain; is_current marks the latest.
Last updated on