Early access
Job Architecture Agent
Benchmarks every role and pay band against the market. Flags drift.
Hours1 (modeled outcome) to re-benchmark a job family, down from 4 to 6 weeks
HR agent
GARecruiting Agent
The Recruiting Agent is a governed AI agent that screens, shortlists, and schedules against the role you defined.
How it works
What the Recruiting Agent does
Scores every application against the requisition's written criteria, produces a ranked shortlist with rationale, and schedules interviews. Recruiters approve every decision.
- 01
Write the criteria once
The hiring manager and recruiter define must-have and nice-to-have criteria on the requisition. The agent uses only these. Criteria are versioned so any score can be traced to the rules in force.
- 02
Screen every applicant
As applications arrive in the ATS, the agent scores each against the criteria, cites the evidence in the résumé or application, and asks knockout questions where the application is silent.
- 03
Recruiter reviews the shortlist
The recruiter sees a ranked list with rationale and an adverse-impact summary. They advance, hold, or reject. The agent drafts the candidate communications for approval.
- 04
Schedule and keep the ATS current
For advanced candidates, the agent finds panel availability, books interviews, sends confirmations, and moves ATS stages. Recruiters see time-to-shortlist and pass-through rates in the Registry.
app.meridian.example/agents/recruiting/req-2291
Senior Payroll Analyst · REQ-2291
25 screened6 shortlisted
| Candidate | Match | Evidence | Decision |
|---|---|---|---|
| A. Okonkwo | 92 | 8 yrs · multi-state payroll | Shortlisted |
| R. Villanueva | 88 | 6 yrs · ADP + Workday migrations | Shortlisted |
| T. Brennan | 81 | 5 yrs · garnishment specialist | Shortlisted |
| M. Aliyev | 64 | 4 yrs · single-state only | Review |
| J. Santos | 41 | No payroll systems experience | Not advanced |
The agent contract
Six terms, applied to this agent
Every Meridian agent runs under the same six terms, enforced at the Gateway below the agent. Here is what each one means for the Recruiting Agent.
Scoped to one workflow
Screens, shortlists, and schedules against one requisition's written criteria.
Reads only permitted data
Reads applications, résumés, and panel calendars. Protected characteristics are excluded.
Every action logged
Every score, rationale, message, and stage change is logged to the requisition.
Consequential actions require human approval
No automated rejections. A recruiter approves every advance, hold, or reject.
Measurable outcome
Time to shortlist and pass-through rates reported per requisition, with the adverse-impact report.
Runs on your chosen model
Choose the model per workspace. Versioned criteria, not the model, decide the score.
Scope
Data, actions, and approvals
What the Recruiting Agent reads, what it does on its own, and where a person signs.
Data it reads
- ATS requisitions, job descriptions, and structured screening criteria
- Applications, résumés, and screening-question responses
- Interviewer and hiring-panel calendars
- Hiring-manager scorecards and interview feedback
- Approved compensation range for the requisition
- Work-authorization and location requirements
Actions it takes
- Score each application against versioned must-have and nice-to-have criteria
- Produce a ranked shortlist with per-candidate rationale and evidence
- Ask knockout and availability questions by email or SMS
- Draft advance, hold, and rejection messages for recruiter approval
- Find panel availability and book interview loops
- Update ATS stages and generate the adverse-impact report per requisition
What needs approval
- No automated rejections: a recruiter approves every advance or reject decision
- Protected characteristics and their proxies are excluded from the features the agent may use
- Adverse-impact monitoring (four-fifths rule) reported per requisition before decisions are released
- Criteria are locked and versioned on the requisition; changes reset scoring
- Candidate data is retained and deleted per the jurisdiction's rules; every score and message is logged
Outcomes
Numbers the Recruiting Agent is accountable for
Modeled outcomes from design-partner deployments. The Registry reports your actuals from the first week an agent is live.
−46%
screening time per requisition1
−70%
manual recruiter reviews1
< 24 h
from posting close to ranked shortlist
100%
of scores with a written rationale
- 1.Modeled outcome from design-partner deployments. ↩
“A dispatcher at 4 a.m. should not be working a phone tree. The agent knows who is certified, who is rested, and who said they want hours. It asks them in that order and books the first yes.”
11min
median time to fill an open shift at Northwind Logistics1
In production at
Integrations
Systems it works with
Connected through prebuilt connectors and zero-copy warehouse access in Data Fabric. The agent acts as the user and inherits each system’s permissions on every read.
- ATS
- HRIS
- Calendar (Google Workspace, Microsoft 365)
- SMS
- Video interviewing
- Background check providers
Credits
What the Recruiting Agent costs
Included on every plan: Starter with 5,000 credits per month, Growth with 30,000, and Enterprise with custom pools. Credits are consumed per completed action, never per message. Usage beyond the plan is billed at $0.12 per credit.
| Action | Credits |
|---|---|
| Candidate screened | 1 |
| Interview loop scheduled | 1 |
| Shortlist package with rationale and adverse-impact report | 3 |
Questions
Questions about the Recruiting Agent
Answered the way buyers ask them. Each answer is also published as structured data.
Related terms
An AI recruiting agent screens applications against the criteria a hiring team defines, produces a ranked shortlist with reasons, and handles scheduling and candidate communication. It does not decide who is hired. It compresses the reading and logistics so recruiters spend their time on the candidates who meet the bar.
It depends on how it is built and where you hire. Meridian's agent uses only the written criteria on the requisition, excludes protected characteristics and known proxies, keeps a human approval on every decision, and produces an adverse-impact report per requisition. Design partners in New York City and the EU have used these artifacts to satisfy audit and notice requirements. Your counsel makes the final call for your jurisdictions.
No. It recommends. A recruiter approves each advance or rejection. Knockout questions such as work authorization can be configured to auto-hold, but the hold is reviewed by a person before any communication goes out.
Any ATS with an API or webhook model. Requisitions, applications, and stage changes sync in both directions. The agent reads the criteria from the requisition and writes rationale back as candidate notes so recruiters keep working in the ATS.
For frontline roles the agent runs knockout questions and availability collection by SMS within minutes of application, then offers interview slots directly. Design partners used it for roles with 1,000-plus applicants per month per location.
Related agents
Agents most often deployed alongside the Recruiting Agent, on the same data and the same approval policy.
Get started
Put the Recruiting Agent to work.
Start with one workflow, one approver, and one number to move. Most design partners were live in five weeks.