Read every CV against the job, not for keywords
Kadroflow's AI reads each CV against the specific requirements of the role and returns a score with evidence for each requirement, rather than matching keywords. Because it judges against the requirements you set, two differently worded CVs are assessed on the same basis, and the reasoning stays attached to the candidate record for the person who reviews it.
Every claim is quoted, or it is dropped
The analysis is citation-verified: a requirement judgement that cannot be backed by a verbatim span of the CV earns no credit, and a run that produces no verifiable citations fails rather than inventing a result. The score you see is traceable to the document, so a reviewer can check the quote instead of trusting a number.
A score is a signal, never the decision
Automatically rejecting, ranking or scoring a candidate as a decision is refused in the system itself — the score is advisory and a person decides and records the job-related reason. Personal-details lines are removed before analysis, and nothing derived from protected characteristics feeds the score, so an advisory signal never becomes a hidden filter.
One pipeline, with the evidence kept
Screening runs per candidate on a queue with a re-run action, and a failed run says what went wrong rather than silently leaving a candidate pending. The score sits on the candidate record beside interview evidence and the team's own review, and AI analyses draw on a shared, plan-based credit budget. Evaluate it with a synthetic CV before using it on live applicants.
Questions before you start
Does the AI reject candidates automatically?
No. Automatic rejection, ranking and scoring as a decision is refused in the system. The AI produces an advisory score with evidence; a person reviews it and makes and records every decision and its job-related reason.
How does AI resume screening work here?
The AI reads a CV against the role's requirements and returns a score with evidence for each requirement. Each piece of evidence must be a verbatim quote from the CV, or it is dropped, and a run with no verifiable quotes fails. The recruiter then reviews the evidence and decides.
Is AI screening fair?
It is built to be a review aid, not an automated gate. Personal-details lines are removed before analysis, inputs derived from protected characteristics are excluded, and the score never rejects anyone on its own. Agreement between the team and the AI is tracked over time on the AI-governance screen so drift is visible.
What does AI screening cost?
Analyses draw on a shared AI-credit budget that comes with your plan, rather than a separate per-CV charge; the assistant draws on the same budget. Current allowances are on the pricing page. There is no charge per candidate for the pipeline itself.