AI for Business
CV Screener against a Job Description
Paste the job description and up to five CVs and get, for each candidate, a fit score with the reasoning, the requirements they meet, the gaps, and the questions to ask at interview. Decision support for a hiring manager: it must not be the only reason a candidate is rejected.
This is a cloud tool. The job description and the CVs are sent to the model. Candidates’ emails and phone numbers are masked on your device first; names, addresses and dates of birth are not, so remove those if you can. Nothing is stored by us; the call is counted against your month’s allowance — 10 free, 300 on Pro. How the AI tools handle data.
Tips
- The score is a starting point for a person, not a sort order for rejection. Read the reasoning and the gaps; a 55 with a career break and the right skills may be the best candidate in the pile.
- Emails and phone numbers are masked on your device before the CVs are sent and restored in the table. Names are not — the model needs them to label candidates — so paste no more than you need.
- The questions to ask column is the interview plan. Verify before offer is the list of claims to check against a reference or a certificate.
- Five CVs a run keeps the notes specific. For a larger pile, run in batches and compare scores across batches with some scepticism: they are relative to the JD, not to each other.
Frequently asked questions
Can I reject candidates on the strength of this?
Not on this alone, and every result says so in its title. It is decision support: a fast, consistent first read of a CV against the requirements, with its reasoning shown so you can disagree. A person should read every CV that is rejected and be able to say why in terms of the job. Automated rejection without human review may also be unlawful where you are — UK GDPR Article 22 and India’s DPDP Act both bear on it.
Is it biased?
Language models can carry biases from their training. The instructions tell it to score on the requirements only and to ignore protected characteristics, career breaks and proxies such as school names; the reasoning column lets you see whether it did. If a judgement looks wrong, it probably is — override it.
What happens to the CVs?
They are sent to the model through our gateway to be read and come back as the table; we store neither. Emails and phone numbers are masked on your device first. If your jurisdiction requires it — the UK does — tell candidates in your privacy notice that an AI tool assists screening.