The prompt
Score 10 of 10, Excellent
Use it when You have a pile of applications and want each checked against the same job-related criteria, with a person making every decision.
- {{role}}
- {{compliance_confirmed}}
- {{criteria}}
- {{applications}}
Full prompt and why it works
Prompt
Help me screen applications for the role of {{role}} against the same must-have criteria, so that I can decide fairly and quickly who moves to a phone screen. I will review every application myself; your job is to show the evidence, not to decide. Has HR or legal confirmed that AI-assisted screening is allowed for this role and location, and that candidates were told where required? {{compliance_confirmed}} Must-have criteria, from the approved job description: """ {{criteria}} """ Applications, each labeled with a reference number only: """ {{applications}} """ For each application, produce one row in a table with the columns Reference, one column per criterion showing "Shown", "Not shown" or "Unclear" followed by the evidence in under 15 words (quoted from the application where possible), and Questions for a phone screen. After the table, list: - Applications where most criteria are "Unclear", which need a closer read rather than a quick no. - Any criterion that screens out almost everyone, so I can check whether it is truly a must-have. Judge only against the criteria listed. Ignore names, photos, age or graduation years, gender, nationality, address, employment gaps, disability, family status and anything else not in the criteria, and do not infer them. Count equivalent experience toward a degree requirement only if the criteria say so. Do not rank or score candidates against each other, and do not reject anyone: "Not shown" means the application does not show it, not that the person lacks it. A good result is one I could explain to a candidate or an auditor, with every mark tied to words in the application. For example, the form of one cell: "Shown: 'managed payroll for 140 staff across two sites' (2021 to 2024)." If the answer to the compliance question above is not a clear yes, stop and tell me to check with HR or legal before screening, because many places now regulate automated screening. If a criterion is too vague to apply consistently, such as "strong communicator", say so before the table and suggest wording an application could show.
Variables
- {{role}}
- The job title.
- {{compliance_confirmed}}
- Write "yes" only if HR or legal has confirmed AI-assisted screening is allowed here and any required notice was given. Otherwise describe what you know.
- {{criteria}}
- The must-have criteria from the approved job description, one per line.
- {{applications}}
- The applications or resumes, each with a reference number. Remove names, contact details and photos first.
Why it works
- What to avoid. It marks evidence per criterion but never ranks, scores or rejects, and it ignores names, ages, gaps and other non-job factors, so the decision stays with a person.
- What good looks like. Every mark must quote the application, the standard you would need if a candidate or an auditor asked why.
- If unsure. Without a clear yes on compliance it stops, and a vague criterion is sent back for rewording before anyone is screened against it.