How to use AI for crafting effective user interview scripts

How can we encourage teams to speak to more users and customers faster? One solution is leveraging AI in the process. Tools like ChatGPT have emerged as valuable assets for crafting user interview scripts. For practitioners in design and product, leveraging these tools can streamline the process, enhance the quality of your scripts, and ultimately lead to deeper insights during user interviews. Here’s how you can effectively integrate AI into your interview script creation process.

Start with a draft using ChatGPT

Begin by inputting your basic interview objectives or any hypotheses you’re exploring and ask it to generate an outline. ChatGPT will suggest a structure for the interview. This will be how you work through the content with AI without having it try to do too much at once, which ultimately leads to poor results.

Build your script like a sandwich

Approach the assembly of your interview script akin to constructing a delicious sandwich. Now that you have the outline—start with one section (ala the bread) and then add each section subsequently (ala cheese, meat, veggies and condiments). Begin with background questions to warm up the interviewee, then move to the meatier, specific questions, and conclude with questions that allow for open feedback, ensuring you gather comprehensive data.

Structuring example:

1. Introduction: Set the tone, explain the purpose of the interview, and build rapport.
2. Background questions: “What do you do for work?”
3. Specific questions: Dive into the details with the questions developed and refined through ChatGPT.
4. Conclusion: “If you had a magic wand and could make 3 changes to the application that would improve its score, what 3 changes would you make? Why?”

Your research intern assistant

Think of ChatGPT as your initial brainstorming partner. ChatGPT can quickly generate a range of questions that cover various aspects of your hypothesis, saving you time and providing a broad base from which to refine your questions. Always ask it to generate more questions than you need so that you can choose the best questions to fit your interview needs.

Example:

If your hypothesis revolves around user satisfaction with a digital tool, ask ChatGPT for open-ended questions that explore user feelings, challenges, and satisfaction levels. This can yield questions like:

• “What was the main job or task you were trying to accomplish the last time you used the app?”
• “Can you describe a time when you weren’t able to complete what you wanted to when using the app?”
• “What were you expecting to happen?”

Refine and expand

Utilize the initial output as a springboard. Review the questions and identify areas that need deeper exploration. You can then request ChatGPT to expand on specific questions or provide variations. This step ensures your script is comprehensive and tailored to gather as much valuable information as possible.

Example:

From the initial question list, you might find that understanding the context of use is crucial. Ask ChatGPT to generate additional questions that dig deeper, such as:

• “Can you generate another 10 open-ended questions that aim to learn more about the struggles they have with their job?”

Generate generic introduction/filler scripts

There is often filler copy in an interview script that lays the foundation for the interview. Creating an effective introduction for your user interview script is pivotal in setting the right tone for the session. Here’s how you can use ChatGPT to ensure your introduction is clear, welcoming, and informative:

Steps to generate an effective introduction:

1. Define the purpose: Inform ChatGPT about the main goal of your interview to tailor the introduction accordingly.
2. Request a welcoming tone: Ask for a friendly and professional tone to put the interviewee at ease from the start.
3. Include essential details: Make sure to include:
   • A thank-you note to acknowledge the participant’s time.
   • Confidentiality assurance if responses are private.
   • An outline of the interview, explaining what to expect and the estimated duration.

Example introduction from ChatGPT:

“Thank you for coming in today! I’m really excited to talk to you today.”

“My name is <moderator_first_name> and I’m a researcher with [company]. This is my research teammate,<observer_name>. She’ll be listening and taking notes. 

I want to make it clear right away that we’re testing the product, not you. You can’t do anything wrong here.

We want to hear exactly what you think, so please don’t worry that you’re going to hurt our feelings. We’re here to improve the experience, so we need to know honestly what you think.

As we go along, I’m going to ask you some questions and ask you to think out loud, to tell me what’s going through your mind. This will help us. Please don’t hold anything back, let us know what you are thinking, and why you are thinking that. 

“I’ll start this session by asking some background questions. I’ll then ask about your particular experiences with [tool/service]. I will then share with you some work-in-progress screens to get your initial reaction. Then we’ll wrap up with final questions after that.”

“Do you have any questions before we begin?”

“Before we begin, we would like to record this session to help us digitize our notes after the fact. Are you comfortable with us recording your voice? It will be anonymous and not shared with anyone outside of our team.” 

Okay great! Let’s get started.

Validate and iterate

After constructing your interview script, use AI to simulate responses or conduct mock interviews. This practice helps identify any gaps in your questions or parts of the script that might lead to biased answers or misinterpretations.

Conclusion

By integrating AI tools like ChatGPT in the development of user interview scripts, design and product leaders can enhance the efficiency and effectiveness of their user research. This process not only saves time but also ensures that the interviews are robust enough to validate hypotheses and gather actionable insights. Always aim to refine based on feedback and evolve your scripts to fit the changing dynamics of user needs and product development cycles.