How AI interviews work
From a new role to a shortlist your team trusts
AI interviews with QuillHire follow six steps. Your team sets up the role and makes the final decision; QuillHire runs the first-round interviews in between and hands back evidence for every candidate.
Last updated
The process, start to finish
- 1
Add a role
Add the job description. QuillHire reads it, and each candidate’s resume, to decide what to ask, so there is no question set to build or rebuild when the role changes. You keep control of the criteria candidates are scored against.
- 2
Bring candidates in
Your team uploads resumes, candidates apply through a link, or they arrive through an employee referral link. Nothing needs installing on the candidate side.
- 3
Choose the format the role needs
A 15-minute phone screen to filter volume, a 45-minute live video interview to test depth, or both. The two are independent, so they run alone or in any order.
- 4
The interview runs on the candidate’s schedule
Phone screens go out automatically within minutes of a candidate entering QuillHire. Video interviews open from a standard video call link. Nights, weekends and peak hiring weeks are covered without anyone on your team being online.
- 5
Review the evidence
Every interview returns a transcript, scores against your criteria, and a recommendation with its reasoning. Video interviews add any code the candidate wrote and anti-cheat findings with timestamps.
- 6
Your team decides
QuillHire recommends; it never makes the hiring call and never rejects anyone automatically. Strong candidates move on to your own interviewers with a packet of evidence already attached.
Which format to use
Pick per role. Many teams run only the phone screen for high-volume roles, and go straight to the video interview for senior or specialist ones.
| Format | Length | Channel | Best for | Anti-cheat | Credits |
|---|---|---|---|---|---|
| AI Phone Screener | 15 min | Phone | Filtering high application volume on fit, logistics and core skills | Not included | 1 |
| AI Technical Video Interview | 45 min | Video | Testing depth in any role, with a coding module for developers | Whole session | 2 |
What anti-cheat monitors, and what to do with a flag, is covered in AI interview anti-cheat.
What the AI decides, and what it does not
The AI decides what to ask next. In both formats it follows up on vague or rehearsed answers instead of moving down a fixed list, which is where most of the useful signal comes from.
It does not decide who gets hired, and it does not reject anyone on its own. Scores and recommendations are inputs for a person. Anti-cheat findings are evidence to look at, with the exact moment attached, not a verdict.
Frequently asked
The process
How fast can QuillHire call after an application?
As soon as a candidate is in QuillHire (via upload, your careers page, or a job board flow), the AI can place a 15-minute outbound phone screen within minutes.
What evidence do we get after each format?
Both produce a transcript, structured scores, and a clear recommendation. The video interview additionally returns any anti-cheat findings as timestamped evidence, plus any code the candidate wrote.
Do the two formats have to run in order?
No. They are independent products. Pick what the role needs: a 15-minute phone screen for volume, a 45-minute video interview for depth, or both.
Who is QuillHire built for?
Hiring teams at startups, staffing agencies, IT services firms, and campus recruiters who run interviews at volume. QuillHire operates in both the United States and India.
Security and data
How is candidate data handled?
Candidate contact details, transcripts, and scorecards are used only to run interviews for your hiring workflows. Access is limited to members of your workspace.
Can we delete candidate data?
Yes. Workspace admins can remove candidates and their associated interview artifacts whenever your retention policy requires it.
How is anti-cheat evidence stored?
Anti-cheat findings live alongside the interview record in your workspace, timestamped against the session so a reviewer can see the moment a signal fired. They are reviewable evidence for your team, not an automated verdict.