You describe the role. Agent writes the JD, posts the job, screens 50 applicants, books the interviews. You approve the hire. ~8 hours a week back.
Each candidate is scored against the role's JD and your store's hire history. Score breakdowns are visible — no opaque black box.
| TITLE | LOCATION | ALL APPLICANTS | NEW | INDE… |
|---|---|---|---|---|
| Team Member | (no local opening) | 49 | 0 | aX… |
| Team Member | Downtown Phoenix | 397 | 198 | aX… |
| Assistant Manager · Ricemill (Arrowhead Towne Center) | (no local opening) | 19 | 0 | aX… |
| Ricemill Store Manager | (no local opening) | 16 | 10 | aX… |
From job posting to interview-set, the recruiting workflow is collapsed into one queue. You stay in approval mode.
Browser-driven worker syncs new applicants every 30 minutes during posting windows.
Each resume is scored on experience, schedule fit, language, prior cuisine match, and your store's hire history.
The screener is fully autonomous. Score breakdowns are inspectable — no opaque black box.
Pre-fills the interviewer assignment, availability window, and message in either language. You confirm with one click.
If the funnel cools, Capo flags it and recommends adjustments — pay band, posting language, or schedule.
The recruiting workflow has three layers — and they have different automation levels. We're explicit about which is which.
PASS / MAYBE / FAIL is generated without you in the loop. You see the queue when you wake up.
An automated worker uses Indeed via a browser session. Reliable, but a human signs off on the posted listing.
The agent fills in interviewer, time slot, and message. You hit approve.
No autonomous offers. The system surfaces the top 3–5 candidates per role, ranked. You make the call.