In the name of research (and curiosity... and maybe a little fear), I recently posed as a respondent in an AI-moderated interview. Not as the moderator. Not as the researcher. Just a regular ol’ human participant, sitting at my kitchen counter with coffee in hand, waiting to be wowed.
I wasn’t.
But I was intrigued.
Here’s What Happened
The setup was straightforward: log in, answer a few closed-ended questions, and then start talking out loud to my computer. No person on the other end. Just me and the bot.
Each response I spoke triggered a follow-up question—automatically generated, presumably based on the “context” of my answer.
It was... fine. Nothing offensive. But also nothing magical. Like ordering a meal that looks good on the menu but shows up lukewarm and missing the sauce.
What the Bot Got Right
It was polite. The AI didn’t interrupt, didn’t judge, and didn’t accidentally call me by the wrong name.
It stayed on script. Maybe too much so—but if you’re trying to collect standardized input across dozens (or hundreds) of respondents, this is kind of the dream.
What the Bot Missed
There was no vibe. No rapport, no warmth, no subtle “I hear you” head nod or curious raised eyebrow that invites you to keep going.
It didn’t get me. It didn’t chuckle at my jokes to encourage me to keep talking. You know I hate when people don’t ‘get’ my humor.
The follow-ups lacked finesse. They were sort of like a new intern trying really hard – repeating what I just said and asking me to elaborate.
I know that this was a sample project and maybe not programmed for in-depth analysis, but I don’t really see it being much better than it was.
The Takeaway
If you’re a real-life moderator wondering whether a robot is about to replace you, let me be clear: not anytime soon.
What I experienced was a perfectly functional, moderately intelligent Q&A session. But moderating? Real moderating—the kind that turns an insight into a story, or a half-baked idea into a lightbulb moment—that takes a human.
A human who listens for what’s not said. Who catches tone shifts. Who digs deeper not because it’s Question 7, but because the moment calls for it.
But Don’t Dismiss the Bot Entirely
There is a place for AI in the research workflow. Just not in your seat. Not yet.
Here’s where I see potential:
- Pre-screening: Let bots collect baseline info or warm-up responses to help you recruit better participants.
- Pre-work or homework: Let participants respond to a few warm-up questions in their own time.
- Internal interviews or training: For companies doing 50 stakeholder interviews, a bot might get you 80% of the way there.
- Large scale saples: If you need 100 people to answer the same 5 questions, the bot can batch that out like a champ.
- Survey open ends: Built into a quant survey to force respondents to elaborate on those brief answers. This might be especially helpful in customer or employee surveys.
- Internal stakeholder interviews: You don’t always need emotional nuance when collecting executive opinions on messaging pillars.
Final Thought
Testing the platform reminded me: tech is a tool, not a threat. And as a tool, it’s kind of amazing. But tools don’t have instincts. They don’t have empathy. They don’t feel when something is just under the surface and worth sitting in silence for.
So yes, I let a bot interview me.
But if I want to be heard, understood, or truly challenged?
I still want a human.
