Abstract
Researchers today will be the last generation doing things the “old way.” Liam Hickey shares his perspective on the evolution of the role of insights professionals in the age of AI.
AI isn’t eliminating the work of researchers. It’s shifting it
Editor’s note: Liam Hickey is the head of research at Knit. He has more than 20 years of experience leading research and analytics teams in the U.S., Australia and in his native home, New Zealand. Find Hickey on LinkedIn.
We’ve seen a lot of gloom and doom news about market research being ripe for job displacement (“Labor market impacts of AI,” “Can AI replace humans for market research?” and “How will AI affect the US labor market?“).
On the other side of the coin, many researchers seem complacent about the impact of AI on their world. At a conference recently, I heard a few research colleagues mention things like, “I put the data in ChatGPT and the numbers were wrong” or “I’m not worried about it – AI can’t do math” or “I’ll be fine. I work in healthcare, finance, government, etc., so we can’t really put our data into AI.” If you also think this, I am here to tell you that this is a critical miscalculation. The fundamental truth is that AI is upending research. And like some of the people at this conference, you can pretend AI can’t do your job, or you can read this article to get a glimpse of how we can shape what the future of research looks like.
What we know to be true today is that people across your organization are DIY-ing their own market research, using AI to fill their customer knowledge gaps. They are asking basic questions: What does [audience] think about [topic]? What are the latest trends in [industry]?
Your internal stakeholders are almost certainly feeding your latest research into the company ChatGPT and asking, “What are the implications for my next [project]?” And just like that, your findings are flattened and reduced to a single user query with an LLM.
Data from a recent Knit study (registration required) among insights professionals looking at how their organizations’ leverage AI is stark: 47% say AI-generated insights are reaching senior leaders with zero researcher review. Another 46% report that non-researchers are using AI to generate data that informs business decisions, and 46% say AI data is being used for strategic decisions without any methodological vetting.
In a positive way, this represents a profound democratization of consumer truth, and an empowering of your business stakeholders. But without the rigor and guardrails of your research expertise, the downside of this environment is a vastly heightened risk of bad future decisions.
If the status quo where all research still flows through the research team is unsustainable, and the other end of the spectrum – where anyone can spin up a “study” – is a nightmare, what is the middle ground?
AI is collapsing the access barrier
Consider this analogy: What AI is doing to research is kind of what WebMD did to doctors.
Overnight, the vast information asymmetry between doctors and patients diminished. Anyone could access detailed medical information and use it how they wanted – for better or worse. It changed the way doctors needed to interact with their patients. But it didn’t negate the role of the doctor to skillfully take in information, make diagnostic decisions and deliver effective treatment recommendations.
AI is doing the same thing to research. It’s collapsing the access barrier. Your stakeholders don’t need to come to you to get an answer anymore. But that doesn’t mean the answer they find on their own is right, nor does it make your expertise less valuable – it actually makes it more critical. Here’s the telling signal from the study mentioned earlier: When asked to choose their bigger concern, researchers picked governance risk over job replacement by more than 2 to 1: 45% vs. 21%. The profession isn’t primarily afraid of being automated out of a job. It’s afraid of being bypassed. Someone in your organization has to make sure the information being used to diagnose and make decisions is sound.
That someone is you. And your new title is Custodian of Consumer Truth.
Researchers have always been trained to do the work but hungered to do the more meaningful work. We’ve been bogged down in designing the questionnaire, cleaning the data, writing the report. We deliver the answers and hope for the best. That’s what the job looked like for decades, and it’s what most of us were hired to do. But that was the execution layer sitting on top of the real job, which is understanding what question is worth asking, whether the methodology is sound, whether the insight is real or an artifact of how the study was designed – and what it actually means for the business.
AI is stripping away the execution layer. And what’s left is the work that only a researcher can do. The evidence shows 90% of insights professionals say AI outputs require editing or rework before they can be used – and the average researcher spends seven hours a week closing that gap manually. AI isn’t eliminating the work. It’s shifting it. The execution layer is getting faster; the judgment layer still requires a human. And that distinction is exactly where the researcher’s value lives.
The Custodian of Consumer Truth doesn’t run every study. They advise how studies are run. They write the rulebook, set the standards and ensure best practices are adhered to. They’re not a gatekeeper, but they verify that effective and innovative research happens within the guardrails of good. They make sure that speed doesn’t come at the cost of quality. When there are competing insights viewpoints, they are the arbiter of what is right. They’re still the person who looks across five years of research and sees the pattern no one else has the context to see. They get pulled into strategy conversations earlier and more often, because they’re not stuck in the weeds, and the business has learned that decisions made without them tend to go wrong.
The profession has already arrived at this norm collectively: 99% of insights professionals agree that researchers need to validate AI-generated insights before decision makers see them. That’s not a niche view held by skeptics – it’s a near-unanimous professional standard. And here’s the encouraging flip side: 96% say that when they are the ones directing the AI, it feels like a tool rather than a threat. The anxiety isn’t about AI itself. It’s about AI without research discipline. Add that discipline back and the tool becomes an accelerant.
We have an opportunity to redefine the role of the researcher to be the Custodian of Consumer Truth. The steward of what good research looks like, what healthy data looks like and what key insights are the headline of the story. The researchers who define this transition will shape what the function looks like for the next decade. The infrastructure is catching up too – research operating systems like are being built for exactly this future. But the tools only matter if you’re willing to embrace the new realities of the role.
We’re not going to be “the last researchers,” but we are going to be the last researchers doing things the old way.
