AI Makes UX Research Faster, Not Optional

AI has changed how quickly product teams can explore ideas.

You can summarize ten papers in minutes.
You can cluster interview notes in seconds.
You can turn a long research session into a structured set of themes within seconds.

And, of course, you can ask AI to design a screen before you have fully understood the problem.

But what happens to UX research when the process gets this much faster?

At UX Camp Europe 2026 in Berlin, I attended a presentation called Using Psychology to Ask Better Research Questions by Van Vo, a User Researcher at AVIV Group.

It was about psychology and how it can help us ask better research questions. But for me, there was a bigger idea behind it. AI is changing the way we approach product design research.

And the main takeaway I took from it is quite simple:

AI can make research much faster. But it makes research thinking even more important.

Let me explain.

Start with the problem, not the interface

The example from the presentation was a team working on a Notification Center.

A very familiar product situation.

There are users with certain problems. There are stakeholders who see the solution differently. There are existing assumptions, pieces of knowledge, hypotheses, and opinions scattered across the team. The challenge is to bring all of this together.

Everyone needs to understand what exactly the team is researching, why it matters, and how the results could affect the product and the design.

This is where I think many product teams can easily jump ahead.

You have a problem. You have Figma. And now you have AI that can generate a dozen interface ideas in seconds. It is very tempting to start there.

But before thinking about the screen, we need to understand the user behavior behind it.

How do users interact with notifications?

What makes them pay attention?

Which triggers actually work?

When does a user need to be interrupted?

When is it enough to simply provide information?

These questions take us back to something much more fundamental: human behavior.

Understand the science of behavior first

I think this is especially important now.

AI can generate a solution super quickly. But if you don't understand how the solution should work, it becomes difficult to make sense of what AI gives you.

You may not notice that something is based on the wrong assumption. You may not know what to ask AI next. And you may struggle to write a good prompt because you haven't fully understood the problem yourself.

This is why I still see UX research as one of the foundations of good product design.

Before searching for a solution, it is useful to understand what we already know about people. That can come from user research, behavioral science, previous experiments, or academic research. It can also come from what your own product data tells you.

The point is to understand the problem before asking for the solution.

AI can help you get to the right research faster

In the Notification Center example, Gemini AI was one of the first tools used in the research process.

You might expect Gemini to answer the research question or come up with the final UX solution. In this case, it helped the team get started, find useful keywords, and explore the topic from different angles before going deeper.

I really liked this approach.

Because sometimes the first challenge in research is simply knowing what to search for.

A product team might say: "We want to understand why users ignore notifications."

But research around this topic may use completely different terms to describe the same behavior.

AI can help you broaden the search. It can suggest related concepts, surface terms you may not have thought of, and give you a few new directions to explore.

This is one of the most practical uses of AI in UX research for me.

AI can help move through the research faster. The researcher still decides where to go next.

Then go to the sources

After identifying the relevant topics, the next step was Google Scholar.

And I think this is an important part of the process to highlight.

If you are looking for scientific research, academic papers, or books, Google Scholar is a much better place to start than a regular Google search.

There is nothing wrong with reading Medium articles, product blogs, or LinkedIn posts. I read them too. They can be great for finding ideas. But when a design decision depends on understanding human behavior, I want to know where those claims come from.

A popular article can tell you what someone thinks. A research paper can help you understand what has actually been studied. For me, this is an important part of evidence-based design.

AI can help you find the evidence but it cannot decide how much you should trust it. That still requires research judgment.

NotebookLM as a research knowledge base

Once the materials were collected, Van Vo uploaded them into NotebookLM and used them as a knowledge base.

I really liked this part.

Why use NotebookLM instead of simply asking ChatGPT, Claude, or another AI tool?

Because the way these tools work can be different.

NotebookLM can work as a knowledge base built around the sources you provide. You upload your articles, books, and research papers. Then you can ask questions based on those materials. It can summarize information, connect ideas and, importantly, it can point you back to the source.

For research, that last part is especially valuable because it becomes much more useful when the path back to the evidence stays visible.

Research knowledge should be reusable

This also made me think about something else.

A lot of research knowledge gets lost inside product teams. Someone did interviews three months ago. Someone found an interesting paper. Someone discovered an important behavioral pattern. Someone tested a design hypothesis. The information exists somewhere. But it may be buried in a presentation, a document, or someone's notes.

And then, six months later, another designer faces a similar problem.

"Didn't we research this already?"

Maybe.

But finding that research can take longer than doing a quick new search.

A structured research repository changes this. Think of it as knowledge bites.

After researching a behavioral pattern, you could create a small card with:

• a short summary of the behavior
• the main research insight
• links to the original studies
• relevant examples
• an explanation of how the pattern applies to your product

Now research becomes easier to use.

A designer can refer to it while working on a feature. A Product Manager can use it when discussing a decision. A researcher can connect it to future research. A stakeholder can understand why a particular design direction was chosen. Over time, these insights can also strengthen the UX strategy.

The research becomes part of the team's shared knowledge.

From "I think this is a good solution" to "Here is why"

This is probably one of my favorite parts of this approach.

Imagine presenting a design to stakeholders.

You can say:

"This feels like a good solution."

Or you can say:

"We chose this pattern because of what we found in our research. Here are the behavioral principles behind it, here are the studies that support them, and here is how we applied them to this particular interface."

These are very different conversations.

The second one creates context. It gives stakeholders something concrete to discuss. It helps the team understand the reasoning behind a design decision. And it makes the design easier to explain.

This is especially useful when many people are involved in a product.

Good research can help everyone get on the same page.

Research is still research

There is a quote that has been my personal research motto for a long time:

"Research is what I'm doing when I do not know what I'm doing."

I still believe that.

Research exists because we don't know something yet.

We have a question. We have assumptions. We have different opinions. We have incomplete information. And we need to understand the situation better before making a decision.

AI can help us get there faster.

Great. I'm all for it.

But faster does not mean that we can skip the research itself. It also doesn't mean that we can ask AI one question, take the answer, and move on. And most importantly, it doesn't remove the uncertainty that made research necessary in the first place.

The quality of the output depends on what you bring in

There is another thing about AI that becomes very obvious when you use it for research.

The result depends on the context you provide. If you don't know what you are looking for, AI can give you a huge amount of information without helping you move forward.

If you have already studied the problem, understand the users, know the relevant context, and have good sources, AI becomes much more useful.

You can ask better questions.

You can write better prompts.

You can challenge the answers.

You can spot gaps.

You can ask for connections between different pieces of information.

You can decide what deserves more investigation.

So the relationship between human expertise and AI becomes quite interesting.

The more you know, the more possibilities AI gives you.

AI expands the UX research toolset

AI-assisted research gives us more ways to explore a problem, work with information, and make research knowledge easier to use. It expands the researcher's toolset.

Gemini can help identify keywords and research directions.

Google Scholar can help find academic evidence.

NotebookLM can help organize a body of research and make it easier to work with.

AI can help summarize and structure information.

A research repository can turn scattered findings into shared product knowledge.

And the researcher connects all of this to the actual users, product, and context.

Each tool has a role and the value comes if you know how to combine them.

And the more you understand as a researcher, designer, or product person, the more powerful your AI toolset becomes.

AI research tools can make the process faster and more accessible. It can make research knowledge easier to organize and share. But research still has to happen. And,I think that's a good thing.

Conclusion: AI changes the speed. Research changes the quality.

For me, this is the main takeaway from the UX Camp session.

AI gives product teams a much larger research toolset.

It can help us move faster from a vague problem to relevant concepts, from scattered sources to structured knowledge, and from research findings to usable product insights.

But the quality of that process still depends on human understanding.

You need to understand your users.

You need to understand the context in which they use the product.

You need to know what you are trying to learn.

You need to ask good research questions.

And you need enough product and behavioral knowledge to recognize when an AI-generated answer makes sense.

The better we understand people, the more useful AI becomes as a research tool.

That is why I think AI will make UX research faster. And, perhaps even more importantly, it can make good research easier to integrate into everyday product decisions.

For teams building SaaS, CRM, ERP, notification systems, internal tools, and AI products, that is a capability worth investing in.

Because great product UX still starts with understanding the person on the other side of the interface.

This is also how we approach product design at UITOP. We start with the people, understand the problem, and use research to give every design decision a stronger foundation. If you're working through a similar product challenge, we'd be happy to help.