Onboarding evidence collection with Jev
Checklists already track “watch the video / ship the laptop.” Jev reads “I’m stuck because…” and picks a helper queue (access, equipment, accommodation). It does not mark tasks done and it does not approve legal forms.
This unofficial page is the evidence collection slice of the onboarding exception routing pack. Intent: apply the Jev (TypeSafe System One) decision model to onboarding exception routing evidence collection. Primary search language: Onboarding Jev evidence collection. 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): LMS/HRIS own task state; Jev routes free-text blockers. Not a checklist-clone or HRIS IA photocopy — I-9 and legal holds never Jev-only. Fan-out extra atoms on one request; open a second HTTP call only for a new artifact, not the same state.
Onboarding use-case context
Evidence collection for onboarding exception routing happens before POST /v1/systemone. Jev does not browse your warehouse, retriever, or ESP. You gather the new-hire or customer onboarding note facts, filter them, then ask snap questions. This slice is where fan-out cost math belongs: batch questions, do not re-send state.
Hub: Use cases. Compare, when the other tool is the real job: onboarding checklists.
Evidence Collection inputs
Collect:
- The free-text note
- Two or three checklist statuses (not the whole graph)
- The routing policy excerpt
Never send:
- I-9 / tax document images
- Asking Jev to mark LMS complete
- Medical details beyond what the person already wrote (minimize)
Shape the payload like this once the gather step finishes:
{
"note": { "id": "ONB-55", "text": "I finished payroll forms at my last employer. Also I cannot sign in to the IdP on a screen reader." },
"checklist": { "idp": "blocked", "laptop": "shipped" },
"policy": { "access": "IdP or assistive-tech blockers go to IT-access, not LMS nudges." }
}Decision signals and actions
Each evidence field should change a named answer:
| Id | Type | Job |
|---|---|---|
blocker |
Choice | access / equipment / training / accommodation / already_done / other |
stuck_severity |
Score | How blocked is the person vs a simple FAQ? |
needs_human |
Noul | Should a coordinator reply (vs an LMS nudge)? |
blocker + severity + needs_human in one call. Accommodation should still go to a human even if confidence is high — that conjunction is your code.
Do not treat a Noul of 0.5 as a “medium” onboarding exception routing score — it means yes and no are equally likely. Conjunctions stay in your code.
Guardrails and escalation
If the gather step fails (empty new-hire or customer onboarding note, redaction stripped everything, retriever empty), fail closed on provisioning access or closing a legal form. Do not invent evidence so Jev has something to say. TypeSafe’s confidence-gated examples use a lower bar for recoverable reads than for irreversible actions. Those numbers are illustrations. For onboarding exception routing, treat provision_access as the high bar (provisioning access or closing a legal form). Tune on labels — see offline evaluation.
Evaluation and rollout notes
Your eval set should include thin-evidence cases, not only happy new-hire or customer onboarding notes. Label blocker-class gold from coordinators, plus accommodation gold (always human). 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 | you are here |
| Atomic rules | policy checks |
| Act / review / abstain | confidence thresholds |
| Reviewer payload | human handoff |
| What to persist | audit trail |
| How it breaks | failure modes |
| Labeled replay | evaluation |
| Shadow → canary | production rollout |
FAQ
Should evidence live in the question text?
Put facts in state and point instructions at note.text, policy.access, checklist.idp. Criteria stay stable so you can replay.
When do I split calls? blocker + severity + needs_human in one call. Accommodation should still go to a human even if confidence is high — that conjunction is your code.
Where is the rest of the Onboarding pack? Start with Onboarding input contracts and Onboarding decision workflow. Cluster hub: Use cases.
Employee or customer onboarding? Same pattern: checklist owns tasks; Jev routes leftover language. Keep legal holds out of auto-act.
Can Jev fill the I-9? No. Never Jev-only on statutory forms.
What this page does not claim
- Not an HRIS, LMS, or immigration product.
- No time-to-productivity claims.
- 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.