RemCal Insights × Skaggs: AI Made Us Faster. Humans Made It Right.
- Karina Harutyunyan
- Mar 31
- 4 min read
We gave AI and humans the same research questions. Here's what we found.
Brands are asking a reasonable question right now: if AI can map consumer sentiment, summarize language patterns, and run category analysis in an afternoon, why pay for a room full of people?
It’s a fair question. We wanted an honest answer.
So we ran an experiment.
We put two sets of answers side by side. One came from real people, real conversations. The other came from an AI agent we gave the exact same discussion guide: same questions, same order.
The AI wasn’t wrong. That’s the part that makes this complicated.
The experiment.
REMCAL INSIGHTS -
We had previously run a full round of qualitative consumer research. Real participants, real conversations, a proper discussion guide covering how people relate to a product category, a specific brand, and their own behavior within both.
SKAGGS -
We took that same discussion guide, created an AI agent and gave it to it. Same questions, same order. Then we put the outputs side by side and looked hard at what was there and what wasn’t.
What AI got right.
SKAGGS -
The themes were correct. Trust, authenticity, value, impact, service quality… The AI identified all of them in roughly the right order and with roughly the right language. It picked up on the difference between people who engage with the category regularly and those who engage with it occasionally, and understood that those two groups have different expectations.
On the surface level, it was fine. If you needed a general map of what consumers care about, the AI drew an accurate one.
It knew what mattered, but it just couldn't tell you what it felt like for it to matter.
What AI missed.
REMCAL INSIGHTS -
The gap showed up the moment the questions moved from "what do you think" to "why does it feel that way".
Real participants didn't just name trust as a concern. They walked us through the specific moment it broke for them. The item that arrived slightly different from the photos. The seller who stopped responding after payment went through. The customer service reply that felt like it came from a template. Trust wasn't a theme for them, it was a specific memory with a specific feeling attached to it.
Then there was the projective question, where we ask participants to describe the brand as a person. Not everyone connects with it, but when they do, the answers tend to reveal more than everything else.
One participant said: "The foreign exchange student who speaks the language but doesn't always get it right."
That single answer held more strategic clarity than pages of structured feedback. It showed the brand's genuine effort, the gap in cultural fluency and, most importantly, the goodwill that was still there underneath the frustration. The foreign exchange student isn't rejected, they're just not quite landing yet. That distinction matters a lot when you're deciding how much repair work the brand actually needs to do.
The AI answered the same question with an accurate summary of the brand's strengths and weaknesses. No image, no feeling, no character. Correct on the facts, silent on what any of it actually meant to a real person.
There were other things the AI had no way of reaching at all. The way participants described their purchase journey in this category, not as a single session but as something that could be processed over months. The personal systems they had built to manage the uncertainty of buying. The small platform detail, a feature that removed photo backgrounds for visual consistency, that participants noticed as something that had quietly decreased their trust without them ever consciously registering it.
Nobody writes that kind of thing in a review. They say it when someone is actually in the room with them, asking and listening.
What the gap actually means. And what we're doing about it.
SKAGGS -
This isn't a flaw in the technology. It's just the reality of where AI learns from. It's trained on what people have already written after the experience is over, the feeling has passed, and after the messy human part has already resolved itself. The hesitation, the face that changed in a second, the emotion that was never typed out, the moment before someone decides. That stuff happens before anyone opens a browser to leave a review. By the time it becomes data, it’s already been filtered through reflection, language, and the self-editing that comes with typing something publicly. The rawness is gone.
We build campaigns, brand identities, and packaging that has to land in a fraction of a second on a real person who is not thinking about it. When the research underneath that work is missing the emotional layer, you feel it in the output. AI is fluent in conclusions. It has never been inside a decision. And decisions are what we design for.
REMCAL INSIGHTS -
We stopped asking which approach was better and started thinking about the sequence.
AI covers the territory first–fast, comprehensive, directionally sound. Human conversations go deeper into what the data can't explain, into the specific memories and projective moments and quiet observations that only surface when someone feels heard. What comes out of those conversations goes back into the system and back into the creative to be tested at scale.
Twenty-four people can show something genuinely powerful, but it's still twenty-four people. The foreign exchange student metaphor was one of the important insights we'd heard in a long time, and it was one person. That’s the point. You’re not trying to replace scale. You’re trying to earn the insight that makes scale matter.
One question worth sitting with:
In your last research project, did you know what consumers felt, or what they said? Are you certain those were the same thing?
AI is not the problem. Assuming it's enough is.
RemCal Insights conducts qualitative consumer research for brands navigating shifts in perception, positioning, and category. Skaggs is a brand growth partner helping brands define, design, and deliver, from brand strategy and identity to campaigns, packaging, and product launches.
Comments