Your hiring process is getting more applications. Is it getting better evidence?

A practical playbook for AI recruitment, candidate assessment and leadership readiness

A Talent Crunch Playbook

Two hundred to three hundred applications. Within a few days. For one role.

A participant described this during our September Talent Crunch All Hands. The problem was finding time to assess them properly, while recognising some applicants were using the same AI skills employers wanted to hire.

In October, another participant raised the other side of the problem: if AI fluency is a hiring requirement, how do you assess it fairly?

“Employers want people who use AI. Candidates use AI to apply. Recruiters use AI to screen. Somewhere in this process, we still need evidence of who is right for the job.”

Across our September and October calls, practitioners shared approaches to application volume, AI screening, candidate assessment and leadership gaps. This playbook turns those experiences into actions for Talent Acquisition, People and HR teams. The recommendations below build on the discussion. They are starting points to test in your organisation.

1. Confirm candidate interest before investing interview time

An application does not always mean someone has considered your role.

One participant described candidates whose applications looked suitable, but who reached the interview without understanding the position. Salary or role expectations then ended the conversation.

Another participant suggested a short follow-up asking for a relevant example of experience. This adds an opportunity to confirm interest before committing more interview time.

Start with three checks

• Confirm the candidate understands the salary range, location and working arrangement.

• Ask for one example connected to a core responsibility.

• Explain the next stage and the time commitment.

Keep the request proportionate. A lengthy unpaid task creates another burden for applicants. A participant shared their experience: personalising applications still brought little response.

If you ask for more effort, give people clarity about what happens next.

2. Define screening evidence before configuring AI tools

In September, we discussed the risk of accepting criteria an AI tool extracts from a job description without checking them.

In October, a participant described applications matching screening terms while the experience behind those terms appeared less relevant to the role.

A keyword match needs a closer look. It does not establish competence. Equally, polished or AI-assisted writing does not establish dishonesty.

Before configuring screening criteria, agree with the hiring manager

• The outcomes the person needs to deliver.

• The experience essential from day one.

• The skills someone will learn after joining.

• The evidence supporting each requirement.

For a marketing data analyst, for example, the word “marketing” tells you little on its own. Ask for evidence of analysing campaign performance, interpreting customer behaviour or informing a marketing decision.

Then review whether your screening process recognises this evidence. Check examples it advances and examples it overlooks. Record disagreements and adjust the criteria.

“You still own the assessment when software helps you conduct it.”

3. Assess AI fluency through the candidate’s decisions

AI fluency means choosing, using and checking AI appropriately for the work involved. A list of tools gives you limited evidence of this ability.

Participants shared an approach: encourage candidates to use AI for a case study, then discuss how they used it during the presentation.

One participant asks what candidates have built and what they are proud of. Another focuses on how people define a problem, plan a workflow, self-correct and decide when human involvement is needed.

Combine these approaches in a role-relevant assessment

Ask:

1. What problem were you trying to solve?

2. Where did you use AI, and why?

3. What did you check before accepting its output?

4. What did you change or reject?

5. Which decisions did you keep with a person?

6. What improved, and how do you know?

Agree the assessment criteria before interviewing. Look for a clear explanation of the problem, appropriate tool use, verification and ownership of the result.

Set expectations by role and seniority. Someone expected to build workflows needs a different assessment from someone learning to use AI in everyday work.

Also ask about access. A participant highlighted candidates who had not been allowed to use AI at work. Limited opportunity to practise needs context during assessment.

4. Match AI hiring requirements with the working environment

I asked a question during October’s call: what access do people receive after we hire them?

If you assess someone’s ability to build AI workflows, then give them access to a single chatbot and little support, your hiring expectations and working environment are misaligned.

Before making AI expertise a requirement, establish

• Which tools people are approved to use.

• Which tasks and data those tools support.

• Who provides training and practical help.

• How employees request access or propose a workflow.

• What improvement you expect in the role.

A participant also raised questions about candidate data and AI screening in September. Bring your privacy and procurement colleagues into these conversations. Ask vendors to explain data handling and how their systems contribute to assessment.

Recruiters need enough understanding to explain their process to candidates. “The system handles it” is a poor answer.

5. Assess leadership readiness before filling a management gap

The same demand for evidence applies to internal promotion.

A participant brought a situation from a 35-person company: an executive had nine direct reports, there was no budget for an external head of department, and junior team members were not ready for promotion.

The group explored options. It did not arrive at a tested solution.

One participant emphasised the different skills leadership requires and suggested access to an experienced internal mentor. Another urged the group to assess interest and readiness before assigning responsibility. They also raised concerns about pressure and competition in a rotation arrangement.

For a similar situation

• Ask who wants to lead, rather than assuming everyone wants promotion.

• Define the responsibilities and skills the team needs.

• Offer supported practice, training and mentoring.

• Agree who owns people management while development takes place.

• Review workload, including work removed from an aspiring manager’s existing role.

The number of direct reports alone does not tell you how demanding the role is. Team experience, support needs and working relationships matter.

“A vacancy creates urgency. It does not create readiness.”

Start with one decision in your hiring process

Choose one live role. Review its screening criteria, the evidence requested from candidates and how interviewers reach a decision.

If AI fluency is required, add a structured discussion about the candidate’s approach. Check whether your organisation provides the access needed to use those skills after joining.

“You do not need another assessment stage by default. You need a clear purpose for each stage already in place.”

Bring your hiring or People challenge to the November call

These discussions started with practitioners bringing a problem they were facing at work. Others shared experience, questioned assumptions and offered approaches worth testing.

The Monthly Crunch is Talent Crunch’s online All Hands for people working across Talent Acquisition, People, Culture and HR. We share community updates, open the hot seat for work challenges and make time for questions and peer exchange.

Bring a problem, an approach you have tested or a question you have not resolved. You are welcome to listen first.

🔗 Join the November Talent Crunch All Hands