Can Generative AI Generate Useful Interview Probes?

Can AI Help You Ask Better Follow-Up Questions During Research Interviews?

You ask a participant to describe an experience using generative AI in teaching. The participant responds, "It was difficult at first, but eventually I became more comfortable with it."

The answer is relevant, but it leaves several questions unanswered. What made the experience difficult? What changed? What does the participant mean by becoming comfortable?

These are opportunities for interview probes, the follow-up questions that help researchers explore what participants mean rather than moving immediately to the next prepared question.

Generative AI can produce dozens of possible probes within seconds. Yet an extensive list of follow-ups is not necessarily useful. Effective probing depends on recognizing which parts of a participant's account deserve further attention and asking questions without imposing an interpretation.

Yes, AI Can Suggest Probes, but the Interviewer Must Respond to the Participant

How to Develop and Use AI-Generated Interview Probes Effectively

What Is an Interview Probe?

An interview probe is a follow-up question or conversational prompt used to obtain additional information about something a participant has already said.

Probes may encourage participants to clarify a term, provide an example, describe an event in greater detail, or explain how they interpreted an experience.

For example, if a participant says, "Our department was not supportive of AI," the interviewer might ask, "Could you tell me what you mean by 'not supportive'?"

This invites the participant to explain their own expression rather than requiring them to accept the interviewer's interpretation of institutional support.

Robinson (2023), in Probing in Qualitative Research Interviews: Theory and Practice, discusses probing as a means of obtaining additional depth and detail through clarification, elaboration, illustration, and explanation of earlier responses.

Probing is therefore not simply asking another question. It involves following an existing account in a way that may reveal meanings or details that were not initially expressed.

Distinguish Main Interview Questions, Planned Prompts, and Responsive Probes

Interview guides often contain main questions and optional follow-ups, but the functions of these elements differ.

Terminology varies across qualitative interviewing traditions, and some authors use "prompt" and "probe" interchangeably. The distinction remains practically useful because a planned follow-up topic is not necessarily a question that should be asked regardless of the participant's response.

AI can help prepare an interview guide, but it cannot know in advance which spontaneous probes will become relevant during an actual conversation.

What Types of Interview Probes Can AI Suggest?

Robinson's (2023) DICE approach distinguishes descriptive detail, idiographic memory, clarifying, and explanatory probes. This provides a more systematic way to think about probing than simply asking AI to generate additional questions.

These examples are illustrative applications of the framework, not a mandatory script. Different probes may be appropriate at different points in the interview, and a single response may invite several possible directions.

The value of the DICE approach is that it encourages researchers to consider what kind of additional information is needed rather than repeatedly asking participants to "elaborate."

Use Clarification Probes to Preserve Participants' Meanings

One of the most useful functions of probing is clarifying language that may have several interpretations.

Suppose an instructor says, "Using AI made me less independent."

The researcher might assume that the participant became overly dependent on AI-generated content. However, the participant may instead mean that institutional rules reduced their freedom to decide how AI could be used.

Asking "How did AI make you dependent on it?" would impose the first interpretation.

A more open probe is "Could you tell me what you mean by becoming less independent?"

DiCicco-Bloom and Crabtree (2006) discuss how repeating participants' own words can encourage clarification without introducing the interviewer's preferred terminology.

AI can generate candidate clarification probes, but researchers should privilege the participant's actual language over an interpretation suggested by the model.

Use Example Probes to Move Beyond General Statements

Participants sometimes provide broad evaluations without describing the events underlying them.

For example, "AI has changed the way I teach" communicates a perception but provides little information about the nature of the change.

A probe such as "Can you describe a particular teaching situation that illustrates that change?" invites a concrete account.

Another participant might say that institutional policies are confusing. Asking for a specific occasion when the policy affected a decision may help the researcher understand what happened and how the participant responded.

These probes can produce more contextual detail, although researchers should not assume that every participant can or wishes to recall a specific event.

Where recall is difficult, an interviewer may invite a more general description or allow the participant to explain the experience in another way.

Use Explanatory Probes Without Demanding a Justification

Explanatory probes explore how participants understand the circumstances or influences surrounding their experiences.

For example, after an instructor describes choosing not to use AI for grading, the researcher might ask, "What considerations influenced that decision?"

This invites the participant's explanation without presuming that the decision was motivated by distrust, ethical concerns, or lack of technical knowledge.

Questions beginning with "Why" can also be appropriate, but their tone and timing matter. "Why didn't you use AI?" may sound accusatory in some contexts, particularly when participants feel that their decisions are being evaluated.

Researchers should consider whether a more conversational formulation would encourage reflection without requiring participants to defend themselves.

Can Silence Function as a Probe?

Yes. Probing does not always require another spoken question.

Allowing a brief silence after a participant finishes speaking may provide an opportunity for further reflection. Attentive nonverbal responses can also encourage continuation without redirecting the account.

DeJonckheere and Vaughn (2019) discuss listening, silence, and planned or spontaneous follow-up questions as important features of semi-structured interviewing.

An interviewer who immediately asks another question may interrupt a participant who was still considering what to say.

AI can recommend allowing space for elaboration, but it cannot determine from a prepared transcript alone whether a particular silence is reflective, uncomfortable, or an indication that the participant has finished.

That judgment depends on the interaction and the participant's preferences.

How Should You Prompt AI to Generate Interview Probes?

A useful prompt provides the main interview question, research purpose, participant population, and intended methodological approach. It should ask for optional probes that respond to plausible types of answers without inventing what participants will say.

This prompt encourages a more selective approach than asking AI to generate ten follow-ups for every question.

Researchers should still evaluate whether the suggested probes are relevant to their study and whether they preserve opportunities for unexpected participant accounts.

Can AI Generate Probes From an Actual Participant Response?

AI can suggest follow-up questions when provided with a participant's response. For example, given the statement "I stopped using AI because it created more work," it might suggest asking what additional work was involved or how the participant reached that decision.

However, there are important limitations.

First, the model may infer motives that the participant did not express. It might assume that the additional work involved checking inaccurate outputs, even though the participant may have been referring to institutional documentation or technical difficulties.

Second, using AI during live interviewing may distract the researcher from attentive listening and affect the conversational relationship.

Third, entering participant statements into an external AI service may raise confidentiality and data-protection concerns, especially when the responses contain identifiable or sensitive information.

Researchers should verify whether such use is consistent with the approved research protocol, participant consent, institutional requirements, and applicable service terms.

For many studies, preparing possible probes in advance and exercising judgment during the interview may be more appropriate than relying on real-time AI suggestions.

How Can Probes Become Leading?

A follow-up question may introduce an interpretation that was not present in the participant's original answer.

Suppose a participant says, "I had some concerns about using AI in class."

A leading probe would be "Were you worried that AI would encourage academic dishonesty?" unless the participant had already introduced that concern or the researcher had a justified reason to explore it as a separate topic.

A more open probe is "What kinds of concerns did you have?"

Similarly, repeatedly probing for negative experiences after a participant has described positive ones may signal that the interviewer expects criticism.

AI can help identify leading interview wording, but researchers must also examine how the timing, repetition, and direction of probes influence the interaction.

Do Not Use Probes to Force Depth Where Participants Do Not Want It

Qualitative interviewing often values detailed accounts, but obtaining more detail is not always appropriate.

A participant may provide a brief answer because the experience is unimportant, difficult to recall, sensitive, or something they do not wish to discuss.

Repeated requests for elaboration can become intrusive, particularly when the topic involves professional conflict, personal distress, or potentially identifiable incidents.

Researchers should respect participants' boundaries and the consent arrangements governing the interview.

AI can help identify potentially sensitive interview questions, but it cannot reliably predict an individual's willingness to discuss a particular experience.

Effective probing involves recognizing when to pursue an account and when to allow the participant to move on.

Why Probes Must Remain Connected to the Research Inquiry

An interesting participant statement does not automatically justify an extended line of questioning.

For example, an instructor may briefly mention dissatisfaction with the university's parking arrangements while discussing preparation for teaching. The topic could be important to the participant, but its relevance to a study of AI-related instructional decision-making may be limited.

Researchers should balance openness to unexpected information with the purpose and scope of the study.

This does not mean dismissing unexpected material simply because it was not anticipated. An apparently peripheral account may reveal an important contextual influence.

The decision requires substantive judgment about the phenomenon being investigated, not merely a rule for maximizing the number of follow-up questions.

Preserving researcher judgment in AI-assisted qualitative question design is therefore especially important during probing.

Should AI-Generated Probes Be Pilot Tested?

Probes included in an interview guide should be reviewed for clarity, relevance, and potential leading assumptions.

Pilot interviewing can reveal whether planned prompts encourage elaboration or make the conversation feel repetitive, overly structured, or intrusive.

Castillo-Montoya's (2016) Interview Protocol Refinement framework includes feedback and piloting as important parts of developing an interview protocol.

However, spontaneous probes cannot all be scripted or tested in advance because they arise from participants' actual responses.

Researchers should therefore combine preparation with interviewer training, reflexivity, and attentive listening.

AI-generated probes may improve preparation, but they cannot replace pilot testing an interview guide when the researcher needs to evaluate how the planned questions function during real interactions.

Using AI to Prepare Probes Without Assuming What a Participant Means

The probes illustrate different ways of exploring one response. They are not intended to be delivered automatically in sequence.

The researcher's task is to recognize which follow-up, if any, is appropriate at each point in the conversation. AI helped prepare possible questions, but the participant's account determines how the interview develops.

Common Mistakes When Using AI-Generated Interview Probes

Prepare Possible Probes, Then Let the Participant's Account Guide Their Use

When preparing an interview guide, use AI to identify a small set of possible probes for each main topic. Consider whether the follow-ups invite clarification, specific experiences, descriptive detail, or the participant's own explanations.

During the interview, prioritize attentive listening over following the prepared list. A useful probe should respond to something meaningful in the participant's account and remain connected to the research inquiry.

Review AI-generated probes before the interview and avoid using confidential participant information in external AI systems without appropriate authorization.

The value of probing lies in helping participants explain what they mean, not in demonstrating how many follow-up questions the interviewer prepared.

Before Using AI-Generated Interview Probes

Frequently Asked Questions About AI-Generated Interview Probes

AI Can Prepare the Probes, but Listening Determines Which Ones Matter

How to Cite This Guide

This guide is intended to be read, shared, and used in research, teaching, and academic work. If you draw on its ideas, explanations, or other content, please acknowledge the source by citing the guide. Doing so gives appropriate credit and helps your readers locate the original resource.