I started my career in market research. Datasets, modelling, conjoint analysis, MaxDiff. The full quantitative toolkit. My job was to turn numbers into decisions for people who needed to understand what their customers actually valued, not just what they claimed to value. It was disciplined work and I enjoyed it. 

Then the world went digital, and the problems changed. The interesting questions were no longer sitting in spreadsheets. They were in the gap between what organisations built and what people actually needed. I followed that shift into user experience, spending the next decade researching and designing services that needed to work for real people. 

The tools I used changed completely. Depth interviews, contextual research, journey mapping, usability testing. I traded regression models for sticky notes, and for good reason. If you want to understand why someone abandons a government service halfway through, a survey will rarely give you the answer. You need to sit with them, watch what they do, and pay attention to where they hesitate. 

I do not regret my move into qualitative work for a second. It remains one of the most powerful tools we have for discovering problems nobody thought to ask about. However, something is shifting again. 

Generative AI has dramatically shortened the gap between intent and syntax. Not long ago, turning an idea into working software was a long chain of translation. A business need would be discussed in strategy meetings, turned into requirements documents, rewritten as user stories, interpreted by designers, handed to engineers and slowly assembled into code. Each step introduced delay, reinterpretation and the occasional misunderstanding that nobody caught until someone tested it with actual users. 

The journey from idea to working product often took months. Sometimes years. That chain is collapsing, and quickly. 

I have watched product managers go from a brief to a clickable demo in the time it used to take to schedule a planning meeting. This changes things for research, and I am not sure we have fully thought through how. 

Software used to be expensive to build. Now it is expensive to misunderstand. For years, user research justified itself as protection against expensive engineering mistakes. We validated ideas before teams committed months of effort to building them. When development cycles were long, the question “are we building the right thing?” had an obvious place in the process. 

When building becomes fast and cheap, that question does not disappear. If anything, it becomes more dangerous to ignore. AI will faithfully scale whatever intent you give it. Clear intent grounded in evidence can produce remarkable results. Vague or mistaken intent gets scaled just as efficiently, and the cost does not vanish, it just moves. 

Instead of wasted engineering effort, you get wasted adoption campaigns, confused users and the slow erosion of trust that happens when services technically function but miss the point in ways that are hard to pin down and harder to fix after launch. 

Where does that leave research? The qualitative depth that user researchers are known for still matters. Understanding people in context, observing real behaviour, uncovering needs that users themselves struggle to put into words. That is not going anywhere. 

I have started to feel that depth on its own is not enough anymore. When teams can spin up new features and variations every week, evidence cannot come solely from occasional rounds of interviews or testing sessions booked four weeks in advance. It has to be gathered continuously, inside live systems, as products evolve. That requires a kind of measurement rigour that is uncommon. 

This is where my earlier career starts to feel relevant again. The quantitative discipline of market research has spent decades building tools for exactly this problem. Experimental design, trade-off modelling, multivariate testing, behavioural measurement. These methods exist to show you what people actually choose when faced with real constraints, not what they tell you in a discussion guide. 

Most user researchers are trained deeply in qualitative methods and only lightly in quantitative ones. That made sense when the job was mostly about discovery and early-stage validation. It makes less sense when your product is changing every sprint and the decisions keep coming whether you have evidence ready or not. 

The profession does not need to abandon qualitative research. That would be a terrible idea, and I say that as someone who has built a career on it. But the next generation of user researchers does need to be comfortable designing experiments, working with behavioural data and measuring real trade-offs alongside the contextual, human-centred work they already do well. 

That means bringing together two traditions that have lived apart for too long. The contextual understanding that user research developed, and the measurement discipline that market research refined. 

I have lived in both worlds, and the gap between them has always struck me as more cultural than intellectual. The goal of research has not changed. It still exists to help organisations understand what people truly value. What has changed is the speed at which those decisions become reality. 

When building becomes easy, intent becomes the real constraint. Research exists to make sure we understand that intent before we scale it.

Chris Leo is principal UX consultant and founder at Pivot UX