| Stage | Definition |
|---|---|
| Acquisition | Target talent acquisition managers and solo recruiters at tech companies with 20–200 employees running active hiring |
| Activation | First successful evaluation completed end-to-end — JD plus PDF to scored result to email draft — within 10 seconds |
| Retention | Recruiters who complete 5+ evaluations in week one return for week two at significantly higher rates due to time saved |
| Revenue | Cost per evaluation under $0.06 at current Gemini API pricing. At 100 evaluations per month total infrastructure cost remains zero on free tiers |
| Referral | A recruiter who screens 20 candidates in the time it previously took to screen 2 has an immediate story to tell their hiring manager |
| Metric | Target |
|---|---|
| Happiness | Recruiter satisfaction score ≥ 4.5/5 |
| Engagement | Average evaluations completed per session; history page return rate |
| Adoption | % of incoming applications processed through Recruit-AI vs manual review |
| Retention | Recruiters active across 2+ consecutive weeks |
| Task Success | Evaluation time from PDF upload to results display consistently under 10 seconds |
| Metric | Target |
|---|---|
| Output Reliability | > 99% valid JSON response rate |
| Hallucination Rate | < 2% |
| Consistency | > 95% reproducibility for identical resume and JD pairs |
| Latency | 5–10 seconds end-to-end including database write |
| Cost per evaluation | < $0.06 at current API pricing |
Screening Speed × Recommendation Accuracy
The system succeeds when recruiters evaluate more candidates per hour without sacrificing decision quality. Both must improve simultaneously.