What Makes a CV Screening Criterion Fair?

A fair CV screening criterion is role-related, evidence-based, consistent, and reviewable.

That sounds simple, but many hiring processes rely on criteria that are vague, informal, or hard to defend:

• "Good culture fit."
• "Top university preferred."
• "Fast-paced background."
• "Strong communicator."
• "Must have worked at a similar company."

Some of these may point to real needs. But as written, they are too loose for consistent CV screening.

The Four Tests

Before using a criterion in AI or manual screening, ask four questions:

Test	Question
Role-related	Does this directly connect to work the person will do?
Evidence-based	Can the CV reasonably show evidence for it?
Consistent	Can every candidate be assessed against the same standard?
Reviewable	Can a human explain why the criterion affected the outcome?

If a criterion fails one of these tests, rewrite it before screening.

Bad Criteria vs Better Criteria

Weak criterion	Problem	Better version
Good culture fit	Vague and subjective	Evidence of cross-functional collaboration
Top university	Prestige proxy	Degree or equivalent evidence where required
Fast-paced background	Ambiguous	Experience handling multiple active client accounts
Strong communicator	Hard to infer consistently	Evidence of stakeholder updates, reports, or demos
Similar company	Can exclude transferable talent	Similar customer type, scale, or workflow

The better versions are not weaker. They are clearer.

Keep Must-Haves Short

Fair screening gets harder when the must-have list is too long.

Use must-haves only for genuine requirements:

• Legal or regulatory requirements.
• Essential licence or certification.
• Required language.
• Location or right-to-work constraint.
• Non-negotiable technical environment.

Everything else should usually be weighted, not mandatory.

That gives strong candidates room to show transferable evidence.

Avoid Prestige Proxies

Prestige signals are tempting because they are easy to scan:

• Famous employers.
• Elite universities.
• Specific job titles.
• Certain career paths.

The problem is that they often stand in for the thing you actually care about.

Instead of "worked at a top SaaS company", ask for:

• Managed a similar customer segment.
• Owned a similar sales cycle.
• Built in a similar technical environment.
• Worked at a similar scale.

That produces a better criterion and a wider candidate pool.

Make AI Criteria Human-Reviewable

AI screening should not hide weak criteria behind automation.

Before processing CVs, a human should review the generated rubric and ask:

• Would we be comfortable explaining this to a candidate?
• Could this exclude people for reasons unrelated to the role?
• Is the evidence rule clear?
• Is this a must-have or just a preference?

If the answer is unclear, rewrite the criterion.

Related Reading

• GDPR-compliant CV screening
• AI CV screening scorecard template
• How to respond when candidates ask about AI CV screening
• Best AI CV screening software UK

---

Want reviewable criteria before candidate data is processed? Marxel turns role briefs into a rubric your team can inspect and adjust first. See how it works