Email confidence thresholds with Jev
A draft exists (human or LLM). Jev judges policy, tone, and whether a promise is unsupported. Code sends, holds, or requests a rewrite. Jev does not deliver mail.
This unofficial page is the confidence thresholds slice of the outbound email decisions pack. Intent: apply the Jev (TypeSafe System One) decision model to outbound email decisions confidence thresholds. Primary search language: Email 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): Send/hold/rewrite as typed decisions + policy excerpts — not a clone of email-automation IA or a drip-builder page.
Email use-case context
Thresholds turn outbound email decisions 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 outbound email decisions hides calibration.
Hub: Use cases. Compare, when the other tool is the real job: email automation.
Confidence Thresholds inputs
You need (1) pinned answers on a frozen contract and (2) labels for send / hold / rewrite gold from brand+legal, plus policy-violation gold. State shape:
{
"draft": { "subject": "Your Q4 discount", "body": "Locked 40% off if you reply today." },
"recipient": { "segment": "trial", "consent_email": true },
"policy": { "promises": "Do not invent discounts not in offer_table.", "tone": "No urgency dark patterns." },
"offer_table": { "trial_discount_pct": 20 }
}Decision signals and actions
| Axis | Where it lives | Email use |
|---|---|---|
choice / score / noul |
answer payload | What to do with the email draft + recipient context |
confidence |
Choice & Score only | Whether to trust the argmax |
| Distance from 0.5 | Noul | Whether policy_ok is decided |
FLOORS = {
"internal_preview": 0.50, # illustrations — replace
"send": 0.88,
}
NOUL_TAU = 0.80 # for policy_ok
def allow(ans, action):
return ans.confidence >= FLOORS[action]Do not treat a Noul of 0.5 as a “medium” outbound email decisions 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 outbound email decisions, treat send as the high bar (sending to a real recipient). Tune on labels — see offline evaluation.
Band around 0.5 on policy_ok always reviews. Do not copy 0.80 onto Choice confidence.
Evaluation and rollout notes
- False-send on a planted bad-discount set
- False-hold (throughput)
- Rewrite rate — if everything rewrites, criteria are mushy
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 send 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 Email pack? Start with Email decision workflow and Email human handoff. Cluster hub: Use cases.
Does this page sell email infrastructure? No. It is an unofficial decision-layer pattern. We do not sell keys or ESP access.
Can Jev write the email? Not in this workflow. Generate elsewhere; Jev gates. That keeps cite/send decisions inspectable.
What this page does not claim
- Not an ESP or CDP.
- No deliverability benchmarks.
- 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.