How to Respond When Candidates Ask About AI CV Screening

Candidates are starting to ask better questions about AI in hiring:

• Was my CV screened by AI?
• Did a person review my application?
• What criteria were used?
• Can I challenge the result?
• Was my data used to train a model?

Recruiters need clear answers. Not defensive answers. Not legalese. Clear, plain-English explanations that show the process is controlled, fair, and reviewable.

This guide gives you a practical response framework.

The Short Answer

When candidates ask about AI CV screening, explain four things:

1. Where AI was used in the process.
2. What criteria were assessed against the role.
3. Whether a human reviewed the recommendation before a decision.
4. How candidate data is handled, including retention and model training.

If your screening tool cannot support those answers, that is a process risk.

Put This in Your Recruitment Privacy Notice

The best time to explain AI screening is before candidates ask.

Your recruitment privacy notice should cover:

Topic	Plain-English explanation
Use of AI	Whether CVs may be reviewed with automated assistance
Purpose	To compare applications against role-related criteria
Human oversight	Whether recruiters or hiring managers review recommendations
Decision logic	The types of criteria considered, not the source code
Candidate rights	How candidates can ask questions or request human review
Data handling	Retention, deletion, processors, and model training boundaries

This is not just a compliance exercise. It also improves trust.

For a deeper legal checklist, see GDPR-compliant CV screening.

A Candidate-Friendly Response Template

Use this as a starting point:

We use AI-assisted screening to help compare CVs against the criteria for this role. The tool looks for evidence related to the job requirements, such as relevant experience, skills, qualifications, and any points that need human review.

The AI does not make final hiring decisions. Recruiters or hiring managers review the output before candidates are moved forward or rejected.

Candidate CVs are processed for recruitment purposes and are not used to train public AI models. You can ask us for more information about the criteria used or request human review of the decision.

Adjust the wording to match your actual process. Do not promise human review if your process does not include it.

What Not to Say

Avoid these answers:

"The AI just ranks everyone objectively."

No screening system is automatically objective. Criteria are chosen by people, job descriptions can contain bias, and model outputs need review.

Better:

The tool applies the same role-specific criteria to every CV, and our team reviews the results before making decisions.

"We cannot explain how it works."

You do not need to explain model architecture. You do need to explain the decision logic in a meaningful way.

Better:

We can explain the criteria used for this role and the evidence that was identified in your application.

"The AI rejected you."

This creates avoidable trust and compliance problems.

Better:

Your application was reviewed against the role criteria. Based on the evidence available, we are not progressing your application at this stage.

What Candidates Usually Want to Know

Most candidates are not asking for a technical audit. They want to know whether the process was fair.

Answer the underlying concern:

Candidate concern	What to explain
"Was I filtered out unfairly?"	Human oversight and role-related criteria
"Did the system miss my experience?"	Route for clarification or review
"Was my data used to train AI?"	Model training and data use boundaries
"Was I compared against biased criteria?"	Criteria review and job-related assessment standards
"Can I appeal?"	Human review or escalation route

Keep an Audit Trail

If you use AI screening, keep records of:

• The job description used.
• The scorecard or rubric confirmed before screening.
• The criteria weights.
• The candidate bucket or recommendation.
• The explanation for each decision.
• Any human changes to the recommendation.

This helps with candidate questions, hiring manager discussions, and internal quality checks.

It also makes AI screening more defensible than unstructured manual review, where the real reason for a decision is often lost in notes, memory, or email threads.

Use Follow-Up Questions for Unclear CVs

Not every weak match is a rejection. Sometimes the CV is simply unclear.

Examples:

• The candidate mentions "leading projects" but not team size or scope.
• Dates suggest a gap, but the context is missing.
• A required tool is not named, but similar experience is present.
• Location or right-to-work status is unclear.

In these cases, a follow-up question is better than an immediate rejection.

Marxel's "Hold" bucket is designed for exactly this situation: the candidate may be viable, but a specific missing detail needs human review or clarification.

Internal Policy Checklist

Before publishing your AI screening wording, confirm:

• Candidates are told AI may assist screening.
• The criteria are related to the role.
• A human reviews recommendations before final decisions.
• Candidates can ask for more information.
• CVs are not used for model training unless explicitly disclosed.
• Retention periods are documented.
• The process can produce a plain-English explanation.

Related Reading

• AI CV screening scorecard template
• Is AI CV screening accurate?
• Manual vs AI resume screening
• Best AI CV screening software UK

---

Need explainable screening outputs? Marxel shows the criteria, evidence, and reasoning behind each candidate bucket so your team can answer questions with confidence. See how it works