AI Workflow Automation
Candidate screening rebuilt so recruiters interview instead of triage
Light-industrial staffing group · 6 branches
The situation
Six branches shared an inbound flow of roughly 1,400 applications a month. Screening was first-come, first-read, so strong candidates aged out while recruiters worked through unqualified volume. Time-to-submit averaged 6.2 days.
What was built
Structured intake with requirement extraction from the req itself; an AI analyst that scores against explicit, documented criteria and drafts a summary with the evidence it used; automatic disqualification only on hard requirements; and a review queue where recruiters confirm or override with one click and a logged reason.
The result
- Time-to-submit down from 6.2 days to 1.8
- Recruiter hours on screening down 63%
- Override rate held at 11%, giving a defensible audit trail on every decision