Five ways recruiters use AI every day (and two they should stop)

Most recruiters already use AI. The question is where it saves an hour and where it quietly costs a hire. From running searches with AI tools inside the pipeline, here is what holds up.

1. First drafts of job posts and outreach

Give the model the role, the team, the three things that make this job different, and your last good post as a style sample. You get a draft in a minute. Then cut everything that sounds generic. The draft is your starting point.

2. Turning a messy intake call into a role profile

Record the call with the hiring manager, paste the transcript, ask for a profile: must-haves, nice-to-haves, deal-breakers, open questions. Send that back to the manager for sign-off before you source. This step alone removes most mis-hires that come from a vague brief.

3. Screening questions and scorecards

From the role profile, ask for five screening questions with what a strong, average and weak answer looks like. Edit them. Use the same card for every candidate. Consistency is what makes your shortlist defensible.

4. Summaries of long interviews

A 45-minute interview becomes a half-page summary in the structure of your scorecard. Read the summary, check it against your own notes, and only then write the feedback for the manager.

5. Research on companies and markets

What a competitor pays, how a team is structured, which companies just had layoffs in your niche. AI is a fast first pass. Verify anything you will repeat to a client.

Two things to stop

Letting AI reject candidates. A model ranking CVs against keywords rejects people for the wrong reasons and you never see them. Use it to surface and sort, never to decline.

Sending AI messages as they come. Candidates recognise them in one line. Keep the structure, rewrite the first sentence yourself, add one specific detail about the person.

The rule

AI is good at the first draft and the first pass. People are good at the decision. Keep the decision.