HiPeople automatically generates screening criteria based on your job description using AI. These criteria help you assess candidates more efficiently and consistently. But you're also in control, you can edit these criteria or create your own to match your exact hiring needs.
When doing so, the prompt behind each criteria is key. A well-written prompt tells the AI what to look for (and what not to), helping it make smart, relevant screening decisions.
What makes a good prompt?
A good prompt:
Uses concrete signals the AI can reasonably detect in a resume
Avoids vague or subjective language (e.g., “good,” “strong,” “top”)
Defines alternative ways a candidate can meet the requirement
Is anchored in reality—think company names, job locations, certifications, years of experience, etc.
Example 1: High English proficiency
✅ Good prompt:
“Candidate either:– Worked 4+ years in a company where the primary language was likely English (e.g., based in the US, UK, Canada, Australia, etc.),– Holds a valid English language certificate (TOEFL, IELTS) from the past 2 years at C1 level or higher,– Was born and raised in an English-speaking country, or– Attended school or university in English.”
This gives the AI multiple paths to infer English proficiency using details commonly found in resumes, like company location, education history, or certifications.
🚫 Bad Prompt:
“Candidate has strong English.”
This is too vague and subjective. What does “strong” mean? There’s no clear signal the AI can evaluate.
Example 2: Worked in top-tier tech
✅ Good prompt:
“Candidate has worked at top-tier tech companies such as Google, Meta, Amazon, Microsoft, Apple, or Stripe. This also includes funded scale-ups—like unicorns, YC alumni with 100+ employees, or companies at Series B+ stage.”
It defines what “top-tier” means by providing company examples and clear alternative qualifiers based on funding and company stage.
🚫 Bad prompt:
“Worked at a good tech company.”
“Good” is entirely subjective. This leaves too much room for interpretation and doesn't tell the AI what to look for.
Working with positive vs. negative criteria
You can define prompts to express what the candidate should have or what they should not have.
Positive prompt example:
“Candidate has experience leading B2B sales teams in the SaaS industry.”
Negative prompt example:
“Candidate has not worked in early-stage startups with fewer than 10 employees.”
Tip: Use "should not" carefully. Only include exclusions that are essential, and make sure they are measurable from resume data.
Binary vs. Optional Criteria
Use binary prompts when the criterion is must-have or must-not-have, and broader prompts when you're open to similar alternatives.
Binary prompt (strict match):
“Candidate has worked at Google.”
Flexible prompt (equivalent experience):
“Candidate has worked at Google or similar top-tier tech companies like Meta, Amazon, Microsoft, etc.”
If you're open to interpretation, make it clear by saying "or similar" and give examples.
What not to use in prompts
To ensure fair and compliant hiring, avoid using personal characteristics that could lead to discrimination:
❌ Gender or pronouns
❌ Age or birth year
❌ Nationality (unless required for legal reasons)
❌ Marital or family status
❌ Appearance or photos
❌ Religious affiliation
✅ Instead, focus on:
Skills
Experience
Certifications
Education
Work history
Achievements
HiPeople is designed to promote fair screening. The clearer and more inclusive your prompts, the better the experience for everyone involved.
If you're ever unsure, start with a clear “if/then” or “either/or” structure. And remember: you can always edit and iterate based on results.
Need help with your criteria? Reach out to your HiPeople CSM rep for support.