Embracing AI-Moderated Interviews: A New Frontier in UX Research
Torn between the scalable quantitative reach of surveys and the depth you get from in-depth interviews? That’s where AI-moderated interviews have really made a difference for me. In this article, we'll explore how AI-moderated interviews bridge the gap between traditional methods, discuss their setup, advantages, limitations, and how they transform raw conversations into actionable insights.
Why choose AI-moderated interviews?
Using platforms like Voicepanel and Outset, we can conduct interviews that maintain the feel of a natural conversation because they're conducted verbally. People actually talk, verbally answering the AI moderators’ questions, which means they tend to share more freely than they might when typing out responses. It's a fantastic way to capture the subtleties of human emotion and expression. See demos of the participant experiences here and here.
AI moderated interviews allow you to conduct shorter interviews (20-30 min appears to be the max for many respondents in my experience) at scale, simultaneously, in multiple languages. It’s a great fit for:
Early on to explore a space and understand the general landscape
Use as a screener for a high-stakes in-depth or in-person interviews
When you want scale, but can’t be as structured as a survey
Later on when you’re digging in on details and collecting examples, or studying the prevalence of a behavior or attitude in different populations/countries
Setting up the conversation
The structure of these AI-moderated sessions should mirror traditional UX interviews: start broad, dig deeper, and finish broad. This format lets us gather comprehensive insights while keeping the conversation flowing naturally. It’s like having a structured chat that gradually peels back layers of the user experience.
It’s important to remember that you’re in control of the AI moderator interview script, and I suggest even when the tool has developed a set of suggested research questions that you carefully evaluate your topic and refine the interview to best meet your needs. In some tools, like Outset, you can specify where you’d like the AI moderator to probe or ask follow-up questions, add skip-logic, etc.
What's the real advantage?
Let’s talk impact. Conducting these interviews in multiple languages? Check. Scalability? Double-check. The verbal format also draws out richer, more candid responses than what you'd typically expect from typed feedback. This isn’t just about cost savings—it’s about enriching our understanding of user experiences at scale.
Conducting interviews in multiple markets/languages is a time consuming and costly process. You may run into issues where different interviewers interpret your study guide differently, or probe on different topics, or simply have differing levels of skill at interviewing. With AI moderated interviews, the same AI moderator is running the interviews across multiple languages meaning you actually gain validity and consistency compared to live interviews because it will probe and follow-up similarly in every interview no matter the language.
When it’s not the right fit
Just like any method, AI moderated studies have their place and aren’t always the right choice.
When more qualitative research is called for: Sometimes you can’t beat in-depth interviews, or contextual interviews when personal rapport or human connection is vital to the topic at hand or it’s about respecting a subject-matter expert's time. Sometimes user interviews can build relationships with potential clients, and this isn’t time to outsource to an AI interviewer! In other cases, immersing yourself or your stakeholders in a real scenario or meeting someone face-to-face is vital to building empathy and understanding. Lastly, sometimes the AI may miss nuances or opportunities to follow-up that you didn’t predict, which can mean missing out on deeper insights so I still suggest conducting some interviews yourself.
When more quantitative research is called for: In other cases, you want something more quantifiable or categorical. This is when it’s more appropriate to ask closed-ended questions in a format like a survey. This is more appropriate when you want to track metrics over time, get counts or percentages, and when you understand the domain well enough to design an appropriate survey.
Turning conversations into insights
Once the interviews are done, the real magic happens in the analysis phase. Most of these AI moderator tools come with summarization & categorization tools that allow you to generate reports quickly.
If you want to go deeper, tools like doReveal and CoLoop are great for digging into the data or comparing cohorts. They help us break down verbal responses and compare different user groups, providing a richer understanding of the data collected.
Conclusion
AI-moderated interviews have proven to be an invaluable tool in my UX research arsenal. They help us maintain a delicate balance between reaching out widely and diving deep where it counts. For anyone straddling the complexities of global user research, this method could be just what you need to gain a deeper understanding of your users.
I am an AI+UX research tools consultant. Feel free to get in touch if you’re curious about how to incorporate AI moderated studies into your work or your team's processes. OR Express interest in my upcoming AI Moderation Course.
