HR & People

AI Hiring & Interviews: Screening to Offer

The Fintra AI hiring flow: create a role, AI candidate screening and scoring, AI voice interviews, transcript and insight review, and converting a hire to onboarding.

Updated July 7, 202610 min read1 labHR AdminOwner / Founder

Hiring at an SMB means the owner reads forty resumes at 10pm. Fintra’s AI hiring stack attacks exactly that hour: screening scores candidates against criteria you define, AI voice interviews conduct structured first-round conversations on the candidate’s schedule, and the hiring copilot assists your human interviews with live context and follow-up questions. Humans make every decision; the AI compresses the funnel between decisions.

Create the role and screen with criteria

The flow starts with a role and its screening criteria - the explicit, written qualities you will score against. Criteria are the control: the AI applies them consistently; you own what they are.

From opening to shortlist

1
Create the role
Define the opening - title, department, must-haves, nice-to-haves. Candidates attach to the role through the recruiting pipeline as applications arrive.
2
Define screening criteria
Write the criteria the screener scores against (e.g., "2+ years field service experience", "valid driver’s license", "customer-facing communication"). Set a minimum score threshold for advancing.
3
Run screening
The screener evaluates each candidate’s materials against the criteria and returns a score with reasoning, plus a recommendation: advance if the score clears your threshold, otherwise reject with the shortfall named.
4
Review the recommendations
A human reviews the scored list - especially the near-misses just under threshold - and confirms who advances to interviews. The screening decision record is kept.

The AI voice interview

Advanced candidates can take a structured AI-led interview - a conversational session (voice-capable, powered by a speech provider such as ElevenLabs) that asks role-specific questions, adapts follow-ups to answers, and produces a transcript with per-answer scoring. It replaces the first-round phone screen, not the human interview.

1Preview the questions the session will ask for the role - adjust before any candidate sees them.
2Create the interview session for the candidate; they complete it on their own schedule, which is half the point for working applicants.
3The session records each answer, asks adaptive follow-ups, and on completion produces the transcript, per-question scores, and a summary.
4Review the rendered transcript and scores in the employer app - read at least the lowest- and highest-scored answers in full rather than trusting the summary alone.

The human interview loop, decision, and hire

Finalists meet humans. Here the AI switches from conductor to copilot: during your interviews it can surface live context (the candidate’s screening and session history), suggest follow-up questions, and summarize answers into structured insights attached to the interview record.

Interview loop: schedule the human rounds; each interviewer’s notes and the copilot’s insights accumulate on the candidate.
Reference checks: run and record them on the candidate before the decision.
Decision: the hiring decision is recorded with its maker. Rejections at any stage should carry the reason - the record is what makes your process defensible.
Hire → onboarding: converting the accepted candidate creates the employee record and starts the onboarding flow (see From Onboarding to Payroll) - the resume-to-first-paycheck path with zero re-keying.

Hands-on labs

Practice against a realistic scenario. Each lab lists the steps, what you should see, and the checkpoints that confirm you got the same result.

Lab 1

Hire a field technician with the AI funnel

Scenario

Acme Services opens a field technician role (the hire budgeted for April). Nineteen applications arrive. You are Jordan, the HR admin; Luis (ops manager) is the hiring manager. Candidate Maria Torres will be your eventual hire.

Steps

1
Create the role with screening criteria: 2+ years field service experience, valid driver’s license, customer-facing communication, HVAC familiarity (nice-to-have). Minimum score to advance: 3 of 5.
Expected: The role exists with four written criteria and a threshold.
2
Run screening across the 19 candidates.
Expected: Each gets a score with reasoning; roughly 6 clear the threshold with "advance" recommendations, the rest show named shortfalls.
3
Review the near-misses (scores just under 3) as a human before accepting the cut.
Expected: You promote one borderline candidate whose resume undersold verifiable experience - the human override is the point.
4
Preview the AI interview questions for the role, adjust one to add a scheduling-conflict scenario, then create sessions for the 7 advancing candidates.
Expected: Candidates complete sessions on their own time; results arrive with transcripts and per-question scores.
5
Review Maria Torres’s transcript - read her top and bottom scored answers in full - and advance her plus two others to the human loop with Luis.
Expected: Interview records show session scores, your review notes, and the advance decision under your name.
6
After Luis’s interviews (with copilot follow-up suggestions and insights recorded) and a reference check, record the hire decision for Maria and convert her to onboarding.
Expected: Maria’s candidate record becomes an employee record with onboarding started - no data re-entered.

Checkpoints - you got it right if…

Screening criteria were written and job-related before any candidate was scored
At least one borderline candidate was human-reviewed rather than auto-cut
An AI interview transcript was read in part, not just its summary
The hire decision is recorded with a named human, and Maria’s onboarding started without re-keying

Frequently asked questions

Can the AI reject candidates automatically?

The screener produces scores and advance/reject recommendations against your threshold, but the process is designed for human confirmation - a person reviews the scored list and owns the cut. Keep it that way: automated rejection without human accountability is both bad hiring and, in a growing number of jurisdictions, a legal problem.

What do candidates experience in the AI voice interview?

A structured conversational session they take on their own schedule: role-specific questions, adaptive follow-ups, and a natural voice interface (speech powered by an integrated provider such as ElevenLabs, depending on deployment). They should be told it is AI-conducted and recorded, with a human alternative available on request.

How is candidate scoring kept fair?

Three mechanisms: criteria are explicit and written by you (so the basis is inspectable), every score comes with reasoning tied to the criteria, and decision records keep who advanced or rejected whom. Audit the funnel periodically - compare score distributions and outcomes across candidate groups - and fix criteria, not symptoms.

Does the AI interview replace human interviews entirely?

No - it replaces the first-round phone screen. Finalists always meet humans; the copilot then assists those human interviews with live context, suggested follow-ups, and structured insight capture, but the interviewers and the decision are human.

What happens to interview data for candidates we do not hire?

Transcripts, scores, and decision records are retained under your HR data retention policy - long enough to defend the process, not forever. Set the retention window deliberately and apply it to AI session data the same as to human interview notes.

Previous in HR & PeopleFrom Onboarding to Payroll: The People Data Flow Next in HR & PeoplePerformance Reviews & the Ombudsman Channel

Keep going

HR & PeopleFrom Onboarding to Payroll: The People Data FlowRead doc HR & PeoplePerformance Reviews & the Ombudsman ChannelRead doc Compliance & TrustAI Governance with AgentFenceRead doc

Go deeper in the guides

AI in AccountingAI Accounting: The Practical Guide for Finance TeamsRead guide

Ready to try it in your own workspace?

Fintra is the AI Finance Operating System for SMBs - accounting, payroll, planning, HR, and compliance under one login, with governed AI doing the heavy lifting.

Talk to us