AI candidate screening
Every profile read against the brief. None of them skimmed.
You define what the role needs. Hyrefeed checks each candidate against every requirement, shows the evidence it found, and marks what it could not confirm, so your first read of a profile starts with the answer.
Sample screened candidate
Three states, always with a reason.
Each requirement lands in one of three states. Met means the profile clearly shows it. Partial means there is related evidence worth a closer look. Not evidenced means the profile does not show it, which is different from the candidate not having it.
- Evidence quoted from the CV, application or referral note
- Not evidenced is flagged as a question for your screen call, not a rejection
- The same brief applies equally to every candidate in the role
Screening result
Priya S.
Software Engineer II, fintech · via Referral
4+ yrs backend engineering
5 yrs 2 mo across two roles
Go in production
Payments service rewrite in Go
Distributed systems
Kafka consumer work, no design ownership stated
On-call ownership
Not evidenced in profile
Pune or hybrid
Based in Pune
Setting up a role
Write the brief the way you would explain it to a colleague.
Start from the job description or a blank page. Hyrefeed suggests structured requirements, and you stay in control of the final list.
Must-haves and nice-to-haves
Mark which requirements are essential and which are a bonus. Nice-to-haves lift a score but never sink one.
Weights you can see
Each requirement carries a weight. Adjust them and the whole feed re-scores immediately.
Edit at any time
Hiring manager changed the brief mid-search? Update it, and every candidate is re-screened against the new version.
Reads the whole profile
CVs, application answers and referral notes are all considered, not just job titles and keywords.
Screen-call prompts
Partial and not-evidenced checks become a ready list of questions for your first conversation.
Job-relevant only
Screen on skills, experience and role needs. Age, gender, religion and similar attributes are never used as criteria.
Questions
Screening FAQ
No. Screening produces checks and evidence. A candidate with several not-evidenced requirements ranks lower, but remains in the feed and can be opened, shortlisted or moved by your team at any time.
Most roles work well with four to eight. Fewer than three gives the ranking little to work with, and more than ten tends to dilute what really matters.
The brief is shared per role, so every candidate is measured the same way. Recruiters can add their own notes and ratings on top without changing the brief.
Thin profiles show more not-evidenced checks. The candidate still appears in the feed, and the gaps are surfaced as questions rather than treated as failures.
Rank 01 is waiting
Put every open role on one ranked feed.
Connect your channels, describe the role, and watch the shortlist assemble itself while you spend time on the conversations that matter.
Sample feed
Re-rankingRiya Kapoor
Matched: 5 yrs SaaS, owns a mid-market quota
Kabir Sethi
Matched: team-lead history, HubSpot admin
Nisha Rao
Partial: 3 yrs SaaS, currently leads a team