Quirk’s London took place May 6–7 at the InterContinental O2, and it draws a different crowd than most industry events. More end-client and brand-side researchers attend Quirk’s than almost any other conference in the industry, which means the conversations in the sessions and the expo hall are grounded in the real problems our industry is facing. And this year, AI in product research was the biggest topic… but probably not in the way you’d expect. Ellie Atack and Tom Vincent recapped their learnings for us below.
AI in Product Research Is Everywhere. Differentiation Is the Real Challenge.
Nobody at Quirk’s was debating whether AI belongs in research. That conversation is over. What filled the sessions and the conversations was something harder: now that everyone has AI, how do you actually differentiate with it?
Speed and scale gains are real. But so are the trade-offs: speed versus depth, automation versus nuance. Many organizations have rolled out AI tools and technologies without matching investment in training, governance, or clear use cases.
One line stuck out to us: AI needs babysitting. Not a knock on the tools, but more of an acknowledgment that adoption without intention produces noise, not a signal.
For you, this means the most dangerous place to be right now is confidently wrong, faster. AI in product research can generate an insight in seconds, but if the research design was sloppy, the sample wasn’t the right consumer, or the question was framed to confirm rather than probe, AI just scaled that mistake.
Bad AI-generated output looks exactly like good AI-generated output. You can’t tell the difference from the report alone. You, and by extension, your partners need to know exactly where the boundaries need to be with AI-enablement.
AI-Moderated Qual: What the Speed Promise is Costing You
One of the most talked-about developments at Quirk’s this year was AI-moderated qualitative and video research. The ability to scale qual without losing depth or to surface the “why” behind consumer behavior in near real-time is a genuine shift we’re all trying to move the needle on. The time-to-insight is compressing in ways that weren’t imaginable three years ago.
But the most important thing said about them was the warning: just because research can move faster doesn’t mean every stage should compress. The speed is in the data collection, but the analysis, synthesis, and interpretation is the real work. The value is still in what a skilled human does with it afterward.
There’s some versions of AI-moderated qual that gives you volume without verdict. You get transcripts and themes fast, but you don’t fully get what you needed. You’re not getting someone who can tell you why that consumer hesitated when she picked up your product. What she almost said before she said something polite. The contradiction between what she told you at minute two and what she revealed at minute twelve. These are patterns that emerge when a trained analyst holds the whole conversation in their head at once. So speed in qual is a feature, but depth is the deliverable. And they’re not the same thing.
What is the Insight-To-Decision Gap?
This landed hard at Quirk’s: insight that doesn’t drive a decision has no value. Full stop. It’s our responsibility to align stakeholders around a clear product story to help drive the product decisions. Insights leaders need to be moving from reporting into recommendation and ensuring the research we do actively shapes product strategy.
One speaker said something that stayed with us: “How can you be agile when the task is to land a plane safely?” Speed is certainly a feature, but confidence is the requirement. In product research especially, the cost of moving fast and landing wrong is measured in launch failures for our clients.
Think about the last product decision that stalled because the team wasn’t aligned on what the research meant. The data itself wasn’t bad, but nobody owned the narrative to decide what it meant. Nobody said “here’s what we know, here’s what we recommend, and here’s why.” This is where product launches stall, reformulations get second-guessed, line extensions die out, and so on. The research existed but it never landed.
Why 96% of Your Data Still Isn't Answering Your Real Question
What product teams need to know about storytelling in consumer research.
This is the number that stopped the room at Quirk’s London: 96% of data tells us what happened. Only 4% explains why. And product leaders don’t act on percentages — they act on narratives. Humans are wired to remember stories, and only the right storytelling can turn a complex body of evidence into a clear, linear narrative that a CMO can walk into a boardroom with and own.
The framing we heard was spot-on: think of yourself as the concert master, not the first chair. The value isn’t in one instrument played well. It’s in bringing everything together into something beautiful that everyone in the room can follow and act on.
Storytelling is the missing link between the data and the impact. The product story is what drives the action, more than data alone can.
The Research Partnerships That Move Products Forward
Every conversation at Quirk’s about AI eventually arrived at the same place: the human at the end of it. It’s the analyst who holds the whole conversation in their head. Or the strategist who pushes back before it goes to field. And the partner who owns the narrative to push the product story forward. The tools are getting faster, but this part isn’t – and shouldn’t.
It’s also critical to know that insight doesn’t and shouldn’t end at delivery. The best research relationships stretch beyond the transactional project. They’re built on someone who knows your product well enough to tell you when you’re asking the wrong questions before it costs you, because we’re also invested in your outcome.