Candidate screening: a 5-step process for small teams

Candidate screening is the work between an application arriving and a real interview: deciding who is worth an hour of your time. The candidate screening process below has five steps: write down what the job needs, ask the knockout questions, send a short async interview, score every answer against the same criteria, and shortlist. You can run it the same way for every role.

The five steps

1. Write down what the job needs, before you read anything

Split it into must-haves (right to work, a license, availability for the shifts) and the three or four things that separate a good hire from an average one. Do this before opening a single application; otherwise the first strong candidate quietly becomes the standard.

2. Ask the knockout questions up front

Put the must-haves in the application itself, as short yes-or-no questions. It is kinder to candidates, who find out in a minute instead of a week, and it saves everyone a wasted interview.

3. Replace the phone screen with a short async interview

Send everyone who passes the knockouts three questions to answer on their own time, by video, voice or text. You review in a few minutes instead of half an hour per call, and every candidate gets the same questions. Our one-way video interview guide covers the setup, and our screening interview questions give you a start.

4. Score every answer against the same criteria

For each question, write one to three criteria for what a good answer covers, and score each from 1 to 5. Note the words that justified the score. A scorecard keeps you honest, and it makes your shortlist easy to explain to a hiring manager or client.

5. Shortlist, decide, and tell everyone

Sort by score, then read the evidence for the people near the line. The score is a starting point; the decision is yours. Then tell everyone where they stand, including the people you are not moving forward. A quick no is better than silence.

How the screening methods compare

Where AI helps, and where it hurts

AI is useful for the parts of screening that are slow but simple:

• Transcribing answers so you can read instead of watch.
• Summarizing the facts a candidate gave, like start date or years in a role.
• Scoring answers against criteria you wrote, and sorting the list so you read the best first.

It causes trouble when:

• It judges things it can't fairly see: faces, voices, accents, gaps in a resume. Several places restrict or ban this.
• It rejects people automatically, so nobody looks at the cases it gets wrong.
• Nobody can see why it gave a score, so nobody can check it.

Our AI candidate screening page shows how we handle each of those. Using AI scores can also create duties for the employer, such as a bias audit in New York City; see our AI hiring laws summary. If you are choosing a tool, our candidate screening software comparison lists eight with prices.

Keep candidates in the loop

Screening is the first time a candidate deals with you, and the people you turn down still talk to friends, apply again, or become clients. Tell them how long the process takes, send reminders before deadlines, and close the loop with everyone. A short, specific no takes a minute to write.