Lead scoring confidence thresholds with Jev
A fuzzy “how hot is this lead 1–10?” prompt hides dimensions. Ask atomic Scores (ICP fit, buying intent, timing), then weight them in code you can change without a new prompt.
This unofficial page is the confidence thresholds slice of the lead scoring pack. Intent: apply the Jev (TypeSafe System One) decision model to lead scoring confidence thresholds. Primary search language: Lead scoring Jev confidence thresholds. Confirm patterns on docs.typesafe.ai. This site does not sell, issue, or proxy TypeSafe keys. Use a credential you already have from the console or a documented gateway.
Independent angle (cover ≠ clone): Weighted Score composition in app code + eval harness + vertical compare — beats a single composite-lead-scoring recipe clone.
Lead scoring use-case context
Thresholds turn lead scoring answers into act / review / abstain. They are product policy, not a hyperparameter TypeSafe ships. Official 0.5 / 0.9 sketches are illustrations. This slice also carries the false-reject discussion: over-gating lead scoring hides calibration.
Hub: Composite scoring. Compare, when the other tool is the real job: classic lead scoring.
Confidence Thresholds inputs
You need (1) pinned answers on a frozen contract and (2) labels for MQL / nurture / recycle gold from SDRs, plus disqualify gold. State shape:
{
"lead": { "id": "L-77", "title": "VP Engineering", "company_size_bucket": "201-500" },
"form": { "message": "Need SOC2 review before Q4 bake-off." },
"firmographics": { "industry": "fintech", "in_icp_list": true }
}Decision signals and actions
| Axis | Where it lives | Lead scoring use |
|---|---|---|
choice / score / noul |
answer payload | What to do with the lead record + recent text |
confidence |
Choice & Score only | Whether to trust the argmax |
| Distance from 0.5 | Noul | Whether disqualify is decided |
FLOORS = {
"nurture_drip": 0.50, # illustrations — replace
"page_ae": 0.85,
}
NOUL_TAU = 0.70 # for disqualify
def allow(ans, action):
return ans.confidence >= FLOORS[action]Do not treat a Noul of 0.5 as a “medium” lead scoring score — it means yes and no are equally likely. Conjunctions stay in your code.
Guardrails and escalation
TypeSafe’s confidence-gated examples use a lower bar for recoverable reads than for irreversible actions. Those numbers are illustrations. For lead scoring, treat page_ae as the high bar (paging an AE or starting outreach). Tune on labels — see offline evaluation.
Band around 0.5 on disqualify always reviews. Do not copy 0.70 onto Choice confidence.
Evaluation and rollout notes
- Precision of AE pages (false pages cost quota)
- Recall of true bake-off language
- Stability of blend after a weight change (replay)
Fit loop: pin jev-1.13.0 → replay → plot error vs confidence → pick floors where auto-act error ≤ your SLA. Pin jev-1.13.0 (the versioned id) after you fit thresholds. jev-latest and the marketing line jev-1.13 can move. Log the response model. TypeSafe’s published list price for jev-1.13 is $0.042 per million input tokens (vendor claim — confirm on the models page); output tokens are free on that same page. Unused distractors still bill as input.
Official Python and JavaScript SDKs read TYPESAFE_API_KEY and retry documented 429/529. This site does not sell, issue, or proxy TypeSafe keys. Use a credential you already have from the console or a documented gateway.
Pack map
| Slice | Page |
|---|---|
| Graph and primitives | decision workflow |
What may enter state |
input contracts |
| What to gather first | evidence collection |
| Atomic rules | policy checks |
| Act / review / abstain | you are here |
| Reviewer payload | human handoff |
| What to persist | audit trail |
| How it breaks | failure modes |
| Labeled replay | evaluation |
| Shadow → canary | production rollout |
FAQ
Should page_ae use 0.9 everywhere? No. Over-gating hides calibration and dumps the queue on humans. Fit per action.
Can I reuse a Noul τ as Choice confidence? No. Jaggedness: they are not interchangeable. See confidence.
Where is the rest of the Lead scoring pack? Start with Lead scoring decision workflow and Lead scoring human handoff. Cluster hub: Use cases.
Why not one Score for “lead quality”? It secretly mixes ICP, intent, and timing. Atomic Scores stay inspectable; weights change in code. See composite scoring.
Can Jev compute a 0–100 predictive score like our vendor? Do not treat a 2–10 rubric as a probability of close. Keep predictive math in your model; use Jev for language judgments.
What this page does not claim
- No invented lift vs Marketo/HubSpot scores.
- Not a CRM.
- Not official TypeSafe.
- Official TypeSafe status, or that jev.pro issues API keys.
- That a schema-constrained answer is automatically factually correct.
Disclaimer
This is an independent unofficial site and is not affiliated with TypeSafe AI; official documentation is available at https://docs.typesafe.ai.
Primary documentation: https://docs.typesafe.ai. Hub: Use cases.
Sources
Public TypeSafe or adjacent documentation only. No private claims.