Use case · People
AI recruitment screening and onboarding, with people deciding.
An autonomous people system assesses applications against the role, shows its evidence, coordinates interviews and runs onboarding from offer to first day. People make every hiring decision.
The problem
Why this work does not scale.
A single open role can attract hundreds of applications, and reading them carefully takes time that hiring managers rarely have. The result is inconsistent screening, slow responses to good candidates and a recruiter's week spent on scheduling and paperwork.
Hiring is also a domain where fairness, explainability and data protection matter a great deal. The right design uses automation to prepare and organize, applies one standard to every candidate and keeps the decision firmly with people.
How it works
From first signal to finished work.
- _01_
Define
Turns the role into explicit, reviewable criteria agreed with the hiring manager.
- _02_
Assess
Reads each application against the criteria and records the evidence from the CV for every rating.
- _03_
Shortlist
Proposes a shortlist with reasons, which a recruiter reviews and adjusts.
- _04_
Coordinate
Prepares briefing packs, aligns calendars and sends interview invitations.
- _05_
Onboard
Runs paperwork, policy questions and first-week tasks from offer to first day.
Authority
You decide how far it goes.
The same function can run at different levels. Most systems begin at Level 2, where people approve each action.
- L1 AssistSummarizes applications and drafts messages. People do the screening.
- L2 ApproveProposes the shortlist and every message. A recruiter approves.
- L3 LimitsHandles scheduling, acknowledgements and policy questions alone. Selection stays with people.
- CeilingHiring decisions stay with a person permanently, by design.
One standard, shown evidence
Every candidate is assessed against the same written criteria, and each rating cites the part of the CV it rests on. That makes screening faster and also easier to review, challenge and correct than an unrecorded first pass by a tired reader.
Data protection and fairness
Applications contain personal data, so access is scoped, retention is defined and every action is recorded, in line with PDPA principles. The system is not used to make or finalize hiring decisions, and criteria are reviewed by your team so they reflect the role and not proxies. This is general information, not legal advice.
Governance
Guardrails built in from the first week.
People decide
The system prepares, ranks with reasons and schedules. Interview outcomes and offers are decided by people.
Reviewable criteria
Criteria are explicit and agreed in advance, so a screening result can always be explained.
Candidate data scoped
Applications are accessed only for the task, kept to retention rules and logged.
Questions
Recruitment screening and onboarding, answered.
Does the AI decide who gets hired?
No. It assesses against criteria, shows its evidence and proposes a shortlist. Interviews and offers are decided by people.
How is bias handled?
Criteria are explicit and reviewed by your team, every rating cites evidence and results can be audited. No system removes the need for human judgement, which is why people decide.
Can it run onboarding too?
Yes. It can collect documents, answer policy questions, assign first-week tasks and keep the new hire and manager informed.
Which languages does it work in?
Thai, English and most major languages, for both applications and candidate communication.
Keep exploring
Related pages.
Professional services
Administration that turns into billable hours.
ArticleAutonomy is a dial, not a switch
Most conversations about AI ask whether to automate something. The better question is how much authority to grant, and how that authority is earned.
GuideSecurity and trust
Workspace isolation, least-privilege access, approvals and audit trails.
Begin
Start with this one function.
Tell Primues where work slows down. It will outline the opportunity and connect you with an architect from our team.