NoRecruit's AI screens every candidate against the criteria you define for the campaign: hard knockouts that exclude candidates outright, and weighted nice-to-have criteria that build a 0–100 score. You can refine criteria after screening and recompute results without re-running the AI.
Where to find it
Criteria live on the campaign's Overview tab, in the Screening Prompt card.
Screening runs from the Candidates tab (Screen Candidates).
Results analysis lives behind the Screening overview button on the Candidates tab.
Building your screening criteria
The fastest way is the Intake Agent: on the Overview tab, add your Job Description first, then click Use Intake Agent. The agent asks a few questions about the role and generates structured screening criteria, saved directly to the campaign.
You can also edit criteria directly in the editor:
Knockouts
Candidates who fail any knockout are automatically excluded (score 0). Knockouts include:
Role-specific criteria — free-text checks, e.g. "Has hands-on React experience".
Experience requirement — Min. relevant years / Max. relevant years, plus an optional Max. total years of experience (regardless of relevance) to exclude overqualified candidates.
Language requirement — a language with a Minimum proficiency (basic / professional / native).
Employment status — Employees only, Freelancers/contractors allowed, or No preference.
Company blacklist — candidates who currently work at listed companies are excluded. Your per-campaign list is merged with the account-wide blacklist, shown as From account defaults (always applied). The account-wide list is managed under Settings → Account Settings → Company Blacklist (see the Settings collection).
💡 Tip: Keep knockouts to 3–4 (usually one being a job-title check). Each extra knockout shrinks your candidate pool — the editor warns you when you add too many.
Nice-to-have criteria
These contribute to the total score. Give each criterion a weight; the weights must add up to exactly 100% before you can save (Distribute evenly balances them for you). Drag criteria to reorder them or move them between the knockout and nice-to-have lists.
Running a screening
On the Candidates tab, select candidates (or none, to target all unscreened profiles) and click Screen Candidates.
Review the dialog: number of Candidates to screen, the Screening Prompt, and the Estimated credits needed — screening costs 1 AI credit per candidate.
Optionally enable Preview on 50 candidates to test your criteria on the first 50 unscreened profiles.
Click Screen X Candidates.
Large runs continue in the background (roughly 30–45 minutes per 1,000 candidates). You can cancel a running screening — unscreened profiles get their credits refunded, and already-screened profiles keep their results.
Understanding scores and results
Each screened candidate gets a score from 0–100 and a Qualified (AI) or Disqualified (AI) badge.
Failing any knockout means Knockout (0). Passing candidates land in score bands: Baseline (60–69), Good (70–89), and Excellent (90–100).
Candidates scoring 60+ count as qualified; with Add qualified candidates to campaign enabled at import time they're added to the campaign automatically.
Click a candidate to open the Evaluation — a per-criterion breakdown of why they passed or failed.
The Screening Overview
Click Screening overview on the Candidates tab to analyze results across the whole campaign:
Overview stats — number Screened, Avg. score, Passed percentage, and a Score distribution chart.
Criteria breakdown — every knockout and nice-to-have criterion with how many candidates passed it (nice-to-haves show full · partial · missing counts). Hover a criterion for its full text.
Excluded candidates — the list of disqualified candidates with the criteria they Did not pass.
The breakdown is interactive: toggle any criterion between Criterion active and Criterion excluded to see a live Preview of the new pass rate and an Estimated score per candidate — without spending credits. When you're happy:
Click Apply to save the changed criteria and recompute the screening results for all candidates in this campaign (runs in the background).
Click Reset to discard your toggles.
💡 Tip: Use the Screening Overview when a knockout turns out to be too strict — exclude it, preview the effect, and apply. The qualified/excluded status recomputes exactly; the score shown in preview is an estimate.
FAQ
Does recomputing in the Screening Overview cost credits? No. Toggling criteria and applying reuses the stored per-criterion results — no new AI calls.
Why is the minimum prompt length 250 characters? Shorter prompts produce unreliable evaluations. The editor and import dialog both enforce this.
Can two campaigns use different criteria for the same candidate? Yes — screening criteria and results are per campaign.