AI might finally deliver on the democratisation of UX research—but only with researchers at the centre.
Editor's note: In this guest post, Naroa Ruiz de Eguilaz, Director of Research Consulting Services at UserTesting, draws on more than a decade of experience in UX research and design to explore how AI could reshape the future of research democratization.
We've been talking about democratising UX research for years. If you've been in the industry long enough, you know the pitch: get UX designers (and sometimes product managers) doing tactical research, so that UX researchers can focus on doing research to influence the big decisions. The discovery work, the strategic research, the complex questions that actually move the needle.
I think it's a good idea, I've always thought so. But from what I've seen, working in consulting and enabling teams on this for years, it hasn't really worked for most organisations.
A lot of people in the UX research community were never convinced it was a good idea to begin with. Many researchers worried that putting research in the hands of designers would lower the quality of the work and lead to poor decisions based on biased findings. There were also deeper fears about expertise being devalued and researchers ultimately being made redundant.
While we were still deciding if democratising was good or bad, the threat of AI arrived. I think that if you put these two threads together, something interesting happens. Two things that each felt like a risk to the profession might, in combination, actually be what finally makes it work. The negative and the negative, if you set it up right, can produce a positive.
ON-DEMAND WEBINAR
Designing the Insight System of Tomorrow: How UXR Leaders Can Shape the Role of Research in the Age of AI
The ambition
The division of responsibilities made sense. Usability testing, copy testing, findability studies, these types of tactical tests don't necessarily require a trained researcher to run. The idea was to let designers own that so researchers could spend their time on the bigger work: understanding users deeply, getting close to them to comprehend their motivations and unmet needs. The kind of research that actually informs what gets built and why.
The problem is what happens in practice.
The challenge to make it work
I've spent years helping organisations set this up, and I've seen the same things arise more often than not. Democratisation tends to be a UX research initiative: researchers want it because they're drowning in validation work, and design leadership often backs it because they want a more user-centric culture. But then it lands on designers who are already stretched, with no OKRs around it, no one asking about it when it's missing, and no real consequence when it quietly disappears from the workflow.
Designers often don't resist the idea, they just don't have the time or the confidence to make it happen consistently. And when they do run studies, the quality is mixed. Not because they're not capable, but because good research is not easy, and it's a skill that takes practice to develop. Before you even get to writing tasks and questions, there are layers of decisions to get right: is there a clear goal, are the research questions well-defined, is the approach right for what you're trying to find out, does the study have the right structure? Each of those takes experience, and when you do research sporadically, as most designers end up doing, it's hard to build that muscle.
I've always made sure the teams we work with receive a solid grounding in the fundamentals: how to avoid bias, how to write good tasks and questions, and how to structure a study properly. But enabling someone once or a few times doesn't make them a researcher. And even when it starts to take off, organisational reality comes along. People leave, teams restructure, new people join the company who might feel differently about the whole democratisation thing. The UX researcher (if they’re still there) ends up back where they started, running tactical studies, still reviewing everything, spending all their time sustaining a process that was supposed to run without them.
GUIDE
Research as Organizational Intelligence : A playbook for research leaders in the AI era
Why AI changes this, and why researchers have to be in the driving seat
Before I jump into how I think AI can change this, I want to be clear about something first: AI on its own is not the answer. Handing designers an AI research tool and stepping back is just a faster way to get the same problems: studies without clear goals that are there just to confirm what the team already believed. The last thing we need is more poor decisions based on more poor data.
What AI does, when a UX researcher or research team is properly orchestrating it, is give researchers a way to scale their judgment. Not replace it, but scale it. The researcher's standards, embedded in a system that touches every study that gets run.
A customised AI, trained on the team's own templates, guidelines and quality standards, becomes the gatekeeper for every study a designer or PM wants to run. The designer brings a research brief, the AI reviews it. If the research questions are too vague it pushes back. If there's no hypothesis, it asks for one. If the methodology doesn't match the question it flags it. And when the brief is solid, an AI trained on research best practice will build a better study than most non-researchers would. Not because it's smarter, but because it's consistent, and because it's been shaped by people who know what good looks like.
This means true governance: standards embedded in the process itself, rather than templates and guideline documents sitting somewhere that no one uses. That's something we've never really had before.
The same applies to analysis. The “marking your own homework” problem has always been one of the strongest arguments against democratisation, and I think it's valid. But AI, trained by a researcher to be a genuinely critical and objective party, can be the thing that breaks that pattern, showing what the data actually says rather than what the team hoped it would.
The researcher's job gets bigger, not smaller
None of this eliminates the need for skilled researchers. The system is not a one-time project, and researchers shouldn’t step away once it’s in place. It’s their job to keep interrogating whether the standards embedded in the AI are still the right ones, to update them when they're not, and to expand what designers and PMs can take on as their confidence and capability grows. The intellectual ownership of how research happens across the organisation stays with them.
The more studies that run through the system, the clearer the picture becomes of where quality still drops and what designers consistently struggle with. And the researcher, who built the system and knows what good looks like, is the right person to read those patterns and act on them.
Another thing that AI doesn’t solve on its own are the organisational problems that have always undermined democratisation efforts: no OKRs, shifting priorities, people leaving, leadership attention moving elsewhere. The AI tackles the quality and speed problem, but it doesn't fix the culture problem. Someone still needs to champion this, and that someone is the researcher. The technology makes their job more scalable, but it doesn't make their presence optional.
This is finally the moment when democratisation might actually work, not because organisations have found the time or the will, but because the technology has caught up with the intention and, for the first time, puts researchers in a position to guide and govern how it happens, not just at the start, but continuously.
If you're a researcher working out how to set this up, this is exactly what we focus on in Professional Services at UserTesting. We've been working with organisations on democratisation initiatives long enough to understand where they succeed and where they struggle, and those lessons are increasingly relevant as teams look to incorporate AI into their research practices.
PODCAST
