Seeing ChatGPT or Copilot on a CV gives you a starting point for a conversation. It does not tell you whether the candidate checks a generated answer before using it, or whether they recognise a task that needs a different approach. If AI is part of the job, those are reasonable things to ask about alongside any specific software experience the role requires.
Here, AI readiness means being able to use the tools relevant to the role responsibly and knowing when to ask for help. The examples below focus on generative AI used for drafting, summarising and research. A job that involves building or managing AI systems will need a more technical assessment.
Decide what the role actually needs
Before adding "AI skills" to a job description, agree with the hiring manager where AI will be used. Someone drafting customer emails needs to understand the product and check that the answer fits the customer's situation. Someone using AI to help analyse financial information needs the relevant numerical knowledge and a way to check the calculations. Those jobs call for different questions, even if both people use the same software.
Specific tool experience can still matter. If a new employee must work in a particular system from their first day, say so and assess the tasks they will need to perform. Where training is available, distinguish what candidates must already know from what they can learn after joining. That keeps the requirements tied to the job rather than to a list of familiar product names.
Give the candidate something to check
A short, role-relevant example can make the discussion more concrete. For an HR role, you might provide a sample employee policy and an AI-generated summary that leaves out an important condition. Ask the candidate to review the summary before it is sent to an employee, with the original policy available for reference.
Listen to how they use the source material and ask them to explain any changes they would make. If they miss something, a follow-up question can help you understand their reasoning, but it should not turn into a different assessment for each candidate. Agree in advance what a satisfactory answer needs to cover and how follow-up questions will be used.
This exercise addresses a documented limitation of generative AI. NIST's Generative AI Profile describes the risk of systems producing confidently stated but false content. A fluent answer therefore needs to be checked against the relevant evidence before it is relied on.
Discuss how much checking the task needs
Finding an error in a sample answer is one part of the discussion. It is also worth asking how the candidate would decide what to check when they are doing the work themselves. An internal brainstorm and a statement about an employee's contractual entitlement should not be handled in the same way. In the second case, the candidate should know which authoritative information to consult and when to involve someone qualified to advise.
For example, ask what they would do if an AI assistant supplied a statistic for a customer presentation without a source. You can then discuss how they would look for the original research and what they would do if they could not verify it. The aim is to understand their proposed process, rather than reward a confident answer about always checking everything.
Ask about a tool they had to learn
Previous experience with a new system can give you something specific to discuss, even if it did not involve AI. Ask the candidate to describe a change they worked through, what they found difficult and how they checked that they were using the system correctly. Follow up on what they actually did rather than accepting a general statement that they learn quickly.
That example is evidence to consider alongside the rest of the interview; it does not prove how they will perform with an unfamiliar AI tool. If a particular skill is essential on entry, include a relevant work sample or technical assessment instead of assuming that experience with another system will transfer.
Talk about where AI use needs limits
Give candidates enough context to discuss the employer's rules as well as the task. For instance, you could ask how they would approach summarising a document containing personal information when they do not know whether the available AI tool is approved for that data. A useful answer would explain what they need to establish before entering the information, rather than assume that every tool is either safe or prohibited.
There are also tasks where a conversation with a colleague, a calculation in an established system or a direct answer to a customer may be more appropriate. Ask the candidate to explain their choice in a situation drawn from the role. Using AI more often should not, by itself, earn a better assessment.
Keep professional knowledge in the assessment
Someone reviewing generated material needs enough knowledge of the subject to question it. A customer support candidate, for example, could be given an answer that promises a refund outside the supplied policy. Discussing the mismatch lets you explore both their understanding of the policy and how they would respond to the customer, without making the exercise depend on experience with a particular chatbot.
Recruitment needs the same attention to the evidence behind a recommendation. In Zeenna's configured screening workflow, candidate answers and supporting information can be returned to the existing ATS for recruiter review. Recruiters retain control over who moves forward. This supports the hiring team's review; it does not make a candidate summary or score a substitute for their judgement.
The interview ideas in this article are general guidance, not a description of a dedicated Zeenna AI-readiness test. Keep the professional requirements of the job in view when deciding how much weight to give any answer about AI use.
Make the questions consistent and relevant
Choose the questions that matter for the vacancy rather than adding a separate AI interview to every hiring process. Use common questions and assessment criteria for candidates applying for the same role, and provide the materials or tools the exercise requires. Explain what the task is intended to assess and make appropriate accessibility arrangements. Candidates should not need their own paid AI subscription to complete an exercise you set.
The US Office of Personnel Management's guidance on structured interviews recommends questions based on job competencies and consistent evaluation criteria. That supports the way an interview is designed; it does not validate the particular examples in this article. Treat these as prompts to adapt with the hiring manager, not as a ready-made test or a guarantee of future performance.
When you review the answers, record the evidence behind your judgement and any gaps that still need exploring. Keep those findings alongside the candidate's professional experience and the other requirements of the role. The hiring decision should reflect the work the person will need to do, including the training and support they will have once they join.