Why AI in UX research has stopped being a debate about replacement

Marc has been asking people that question for years, as the host of his own show UX Pathways. And he's noticed something.

“The conversation has changed,” he said. “It's not, 'Will AI replace UX?' It's more down this practical realm of, 'How do I use it responsibly? Which parts of my workflow should I automate? How do I maintain quality?' Those are more productive questions."

That shift—from existential to operational—is worth sitting with. Two years ago, plenty of researchers were debating whether their jobs would survive the arrival of generative tools. Now the debate has narrowed to workflow design.

“And that's why I think AI is becoming more of a toolkit. It's something that you look at on your tool belt that isn't something that's going to replace everything. It's just merging into how you're going to do that,” he said. “I think really the big takeaway here is that the future is going to belong to those who learn to work with AI, thoughtfully.”

Marc sorts the profession into three camps: the experimenters who use AI constantly, the cautiously curious who dip a toe in, and a third group that mostly just talks about it—attending webinars, sharing articles, never quite integrating it into daily work. He puts himself firmly in the first camp, tracing it back to a childhood spent as an early adopter, "one of those people that got the first iPhone and figured out how to do things."

But even the most enthusiastic adopters, in Marc's telling, are wary of one thing: designing for a machine instead of a person. He raised an unsettling possibility—that just as an earlier generation of websites were optimized to please search engine algorithms, the next generation of products might be built primarily for AI agents, with humans almost as an afterthought.

"That's hard for me to wrap my head around," he admitted, "because I think the human will always be in the loop." He's not so sure everyone else agrees.

Synthetic users, real trade-offs

Synthetic users, sometimes described as "digital twins" of actual users, have become one of the more contested tools on a researcher's toolbelt. Marc doesn't dismiss them, but he doesn't romanticize them either.

"It's not going to replace real users," he said, "but it could be used in very specific scenarios."

His use case: run a test against synthetic users first, adjust the study based on that insight if needed, and then put the test in front of real people. Fewer false starts, more informed questions once the humans show up.

What's notable is what Marc thinks synthetic users are actually good for—not cutting corners, but building appetite. He argues that once a team gets a taste of rapid, low-stakes testing, they tend to want more of it, real users included.

"It's addictive," he said, "it's enjoyable... when in doubt, test it out." 

There's a useful distinction buried in that optimism. Synthetic users, done well, aren't a substitute for curiosity—they're a way of lowering the cost of asking a question, so more questions get asked.

Done poorly, they become a shortcut that lets teams skip the harder, slower work of talking to actual customers. Marc's framework is really a bet that most organizations will use the tool to ask more, not to ask less.

The business case researchers still have to make

None of this matters, of course, if research teams can't survive the budget conversation. Marc has spent 25 years in an industry that has never stopped justifying its own existence, and he's distilled the pitch into four arguments: 

1. Competitive advantage
2. Reduced rework
3. Accessibility
4. A more pleasurable customer experience.

Of those, he says, competitive advantage is the one that lands hardest with senior leaders—even though many researchers are uncomfortable leading with it.

The deeper issue, in his view, isn't messaging. It's timing. He recalled meeting someone who attended the very first Agile conference, who told him that user experience wasn't part of the process at all.

“Let's start over and think of UX early,” he said. “And then account for that time and then start development. That would be better."

It's a modest-sounding fix for a problem that has outlasted every technology cycle it's been asked to survive, including this one. AI hasn't solved the timing problem. If anything, it's made speed more tempting and thoughtfulness more optional.

What holds, even as the tools change

Across two decades, a graduate degree, a wedding-DJ side career, and a podcast full of interviews with researchers, designers, and others, Marc keeps arriving at the same conclusion: the tools change, the mindset doesn't.

Curiosity, empathy, ethics, a willingness to keep learning—that's the throughline he hears no matter who's in the guest chair.

Toward the end of the episode, Rachel asked what single idea he'd want every senior leader to take away. His answer doubled as a closing argument for the whole conversation.

"UX is a different lens," Marc said. "A lot of people think it's about investing in interfaces, but it's really about this lens giving you more insights to make better business decisions." 

It will give you a competitive edge, it will reduce rework, it will allow people to work at their highest ability, and will offer a more pleasurable experience for all. “Those are the metrics that you should be able to gather and wrap your head around,” he said. 

Then he added the mantra, "When in doubt, test it out."

Additional resources

• Research in the age of AI: How to stay relevant, build trust, and prove your worth. This on-demand webinar closely aligns with Marc Majers' discussion about the evolving role of UX researchers, the responsible use of AI, and demonstrating the business value of UX research.
• The state of synthetic users report. Marc discusses synthetic users as a complementary tool rather than a replacement for human research. This report, that surveyed 150 researchers on their use of synthetic users, dives into practitioner perspectives on synthetic users and AI-assisted research.
• AI can build anything. It still needs someone to point the way. This episode explores many of the same themes as Marc's interview, including AI in UX, human judgment, agentic AI, and why people—not technology—should remain at the center of experience design.
• Evolve your UX research strategy to win in an AI world. This on-demand webinar discusses how UX teams combine AI and human expertise to accelerate research and reduce uncertainty, echoing Marc's point that AI is becoming a tool on the belt rather than a replacement for researchers.
• Proving the ROI of UX research. A practical guide for quantifying research impact and securing budget, directly aligned with Marc's four-part framework for making the business case to leadership (competitive advantage, reduced rework, accessibility, and customer satisfaction).

Frequently asked questions (FAQs)