Agent Library · By function · Screening · Onboarding · Utilization
Hiring judgement stays human. The sorting doesn't.
This is the function where careless automation does real harm to real people, so the design is deliberately conservative. Agents here structure, summarise and surface. They do not decide, and they do not rank a human being without a human reading the reasoning.
Screening swallows senior hours
Two hundred CVs for one role, read by the person whose time is worth the most.
Onboarding quality depends on who is free
The same role gets a different first month depending on who happened to have the bandwidth.
Knowledge leaves with people
Everything a departing employee knew that was never written down leaves with them.
The agents
5 agents. Each one earns its autonomy.
Agent
What it does
Autonomy
Human gate
CV-to-Rubric Screener
Scores applications against a rubric you define, showing its reasoning per criterion — never a bare ranking.
Review
Every rejection decision stays human
Onboarding Runner
Runs the structured first-30-days: accounts, documents, reading, check-ins — the same quality regardless of who is free.
Notify
—
Knowledge Capturer
Turns exit interviews, handovers and process walkthroughs into documented, searchable institutional memory.
Review
—
Utilization Tracker
Timesheet chasing, utilization reporting and anomaly flagging — without a manager doing the chasing.
Auto
—
Interview Brief
Assembles the interviewer's brief: what to probe, what the CV leaves ambiguous, what previous rounds surfaced.
Auto
—
Autonomy key — Auto: executes above 95% confidence. Notify: executes at 85–95% with notification and undo. Review: always queued for a human. Flag only: never acts, alerts a person.
People-operations agents carry a lower autonomy ceiling than any other function at Kamaan. No agent in this library makes or communicates an employment decision.
What changes
Three shifts, not a feature list.
Senior time moves from sorting to judging
The pile gets structured and evidenced. The decision stays entirely yours.
Every new hire gets the same first month
Onboarding quality stops being a function of who had time that week.
Institutional memory survives departures
What people know gets captured while they are still there to correct it.
The eval bench
What these agents have to pass before they touch real work.
- Rubric scoring calibrated against 50 applications your team has already assessed, with disagreements reviewed
- Bias check: outputs tested for correlation with name, gender and institution proxies before go-live
- Zero autonomous rejections — verified structurally, not by policy
Written against your data, not ours. The bench is built during Blueprint and runs on every deploy from then on.
Related
Where this connects.
See it on your Google Workspace.
A Blueprint maps which of these agents pays off first in your exact setup — and which to leave alone.
Apply for a Blueprint