AI recruiting agents for frontline hiring
ejob.ai calls your frontline applicants, runs a four-minute interview in English or Spanish, backs every score with a verbatim quote and books the in-person interview.
Alex
AI interviewer · Demo Optical Co.
Incoming call
Optical Sales Associate interview
About 4 minutes · recorded with your consent
Daniel…the insurance approval was stuck, so I called the pharmacist over and we got it approved while he waited.
AlexThank you. In Spanish, how would you greet a customer?
DanielBuenas tardes, bienvenido. ¿En qué le puedo ayudar hoy? Español
Daniel Park
89/100·17 verified quotes
“…so I called the pharmacist over and we got it approved while he waited.”
Phone call
The agent calls. Candidates just answer.
Email link
Interview whenever it suits them.
QR code
Scan in store, start on a phone.
Web video
Face to face with an avatar.
English · Español
Switches language mid-interview.
Product
Each agent owns one step of frontline hiring and hands off to the next. Every hand-off is logged, so you can see exactly what happened with every candidate.
From a one-line brief to a structured interview.
The same structured interview for every candidate.
Scores backed by the candidate's own words.
The next step, already on the calendar.
How it works
Your recruiters set the bar and make the decisions. The agents do the calling, the note-taking and the scheduling in between.
Type a sentence or paste the job description you already use. The Designer drafts the job profile, a weighted scoring rubric and the interview questions.
Job brief
Optical sales associate, San Francisco, bilingual English/Spanish, weekends
Rubric · 7 criteria
DesignerUpload the CVs you have, or add a candidate with just a name and a phone number. Invite the top of the list: the Interviewer calls each candidate, and an email link or an in-store QR code opens the same interview in the browser, by audio or face to face with an avatar.
Right after the interview, each criterion has a 1–5 score with the candidate's own words as evidence, checked against the transcript. The recording and full transcript are one click away.
Candidates who clear your threshold get the in-person interview on your Google Calendar, with invitations. The rest receive a courteous hold email. Every hand-off is timestamped in the agent timeline.
Solutions
Wherever hiring is high-volume, hourly and customer-facing, the rubric adapts to the role and the interview meets candidates on the phone they already carry.
Screen for customer service, weekend availability and reliability before a store manager spends time on a call.
Typical roles
Bilingual counter roles where care with orders, prescriptions and payments matters as much as the sale.
Typical roles
High-volume hiring that keeps up with turnover, with interviews candidates can take between shifts.
Typical roles
Patient-facing roles where calm communication and scheduling accuracy come first, in English or Spanish.
Typical roles
Check shift availability, safety habits and reliability in a short phone interview candidates can take from anywhere.
Typical roles
Hear how candidates actually sound on the phone. For phone roles, the interview doubles as a work sample.
Typical roles
Evidence, fairness and security
Guardrails are part of how ejob.ai works, not a policy page. Recruiters can explain every recommendation, and candidates know what is happening before the first question.
Customer service orientation
4 / 5
“I asked what she needed the glasses for before showing her anything.”
Found verbatim in the transcript at 01:52
Attention to detail
—
Insufficient evidence: not scored
Each interview opens with a recording and AI-evaluation disclosure and continues only with the candidate's consent: spoken on the phone, a click on the web.
Quotes are checked against the transcript in code, not trusted from the model. A criterion without one reads “insufficient evidence”.
The Interviewer never asks about age, family status, health, national origin or other protected characteristics, and the Evaluator never scores them.
Flags are job-related: availability for the schedule, inconsistent employment claims, unprofessional conduct. Never demographic.
Agents recommend advance or hold against the threshold you set. Your people run the in-person interview and decide who is hired.
Every agent hand-off is timestamped in the candidate's timeline, next to the transcript, the recording and the evidence.
Candidate records, transcripts and evaluations live in the database you configure; CVs and recordings in your own object storage.
Traffic is served over HTTPS. CVs and recordings open through short-lived signed links, and phone numbers are masked in logs.
Pricing
Every plan includes all four agents, English and Spanish, consent on every interview and a human in every hiring decision. Pricing is scoped to your hiring volume.
Prove it on one role, at one location.
Scoped pilot
Fixed scope, agreed up front
For teams hiring every week, across roles.
Custom pricing
Based on your interview volume
Everything in Pilot, plus:
For multi-site operators with their own stack.
Talk to sales
Volume and deployment options
Everything in Growth, plus:
ejob.ai is an early-stage product. We will tell you plainly what runs today and what we would build with you.
FAQ
Something else on your mind? Talk to our team.
English and Spanish. The interviewer greets each candidate in their preferred language, and if the candidate answers in the other one, it switches mid-interview and carries on. Your dashboard stays in English.
Every interview opens with a disclosure: the interview is recorded and evaluated with AI to keep it consistent and fair, and a human makes the final decision. On the phone the candidate agrees out loud; on the web they click to agree. If they decline, the interview ends there.
The candidate is marked as no answer and still has the interview link from their invitation email, so they can interview from their phone's browser whenever it suits them. You can call again from the dashboard, and candidates who call back reach the same interviewer.
The Evaluator must support every criterion score with a verbatim quote from the candidate. Code, not the model, checks each quote against the transcript and discards any that don't match. A criterion left without a verified quote is not scored and reads “insufficient evidence”.
No. The agents recommend advance or hold against the threshold you set, and can book the in-person interview for candidates who clear it. Hiring decisions are made by people on your team.
No. On the phone they simply answer. The email link and the in-store QR code open the interview in the browser, with no login, in audio or in video with an avatar.
Today it books interviews on Google Calendar with invitations, sends email from a verified sending domain, places calls over standard SIP telephony and reads CVs as PDF or Word files. Connections to other systems, such as an ATS, are scoped with Enterprise customers.
In your stack: candidate records, transcripts and evaluations in the database you configure, and CVs and recordings in your own object storage, opened through short-lived signed links. Protected characteristics are never written to a profile or an evaluation.
Book a demo
Tell us what you're hiring for. We'll show you the four agents working end to end, on your role, in one short call.
We turn one of your job descriptions into a rubric and run a real interview, by phone and on video.
How disclosure works, what is stored, where it lives and how every score is verified.
One role, one location, a clear scope and the calendar you already use.
