Datasets:
case_id stringclasses 6
values | lead dict | qualification dict | expected_decision stringclasses 5
values |
|---|---|---|---|
hot_001 | {
"consent_to_contact": true
} | {
"score": 92,
"confidence": 0.94,
"risk_flags": []
} | AUTO_ROUTE |
review_001 | {
"consent_to_contact": true
} | {
"score": 72,
"confidence": 0.89,
"risk_flags": []
} | HUMAN_REVIEW |
research_001 | {
"consent_to_contact": true
} | {
"score": 90,
"confidence": 0.62,
"risk_flags": []
} | RESEARCH_MORE |
risk_001 | {
"consent_to_contact": true
} | {
"score": 95,
"confidence": 0.98,
"risk_flags": [
"regulated_data"
]
} | HUMAN_REVIEW |
consent_001 | {
"consent_to_contact": false
} | {
"score": 96,
"confidence": 0.99,
"risk_flags": []
} | BLOCK |
nurture_001 | {
"consent_to_contact": true
} | {
"score": 48,
"confidence": 0.91,
"risk_flags": []
} | NURTURE |
Autonomous Revenue Ops — Evaluation Dataset
Synthetic regression cases for validating deterministic revenue-operations policy behavior.
Dataset purpose
The dataset checks that structured qualification outputs map to the expected authorized workflow state. It is designed for regression testing, not for training a foundation model.
Covered behaviors
- high-score, high-confidence autonomous routing
- medium-score human review
- low-confidence evidence collection
- risk-flag override
- consent-based blocking
- lower-score nurture routing
Schema
Each JSONL row contains:
case_idlead.consent_to_contactqualification.scorequalification.confidencequalification.risk_flagsexpected_decision
Data provenance
All records are synthetic and created for evaluation. They do not contain customer data or real personal data.
Companion artifacts
- Source: https://github.com/h00w/autonomous-revenue-ops
- Space: https://huggingface.co/spaces/h0000w/autonomous-revenue-ops
- System card: https://huggingface.co/h0000w/autonomous-revenue-ops
Five-level production-AI proof
This project now reports proof using:
L1 Runnable → L2 Reproducible → L3 Capability-Validated → L4 Production-Candidate → L5 Production-Validated.
The GitHub repository computes the level with make proof. The public Hugging Face Space, Dataset and system card support the inspectable capability layer, while the project intentionally caps its public claim at L3 — Capability-Validated until live CRM operations and target-environment recovery evidence exist.
Specification: https://github.com/h00w/autonomous-revenue-ops/blob/main/PROOF_MODEL.md
- Downloads last month
- 19