In the Age of AI, User Research Matters More Than Ever

Over the last year, conversations about the future of design have increasingly centered around tools. Designers discuss how they use ChatGPT, Claude, Gemini, Figma Make, UX Pilot, Cursor, and many other AI-powered assistants throughout their workflows.

Yet amid this excitement, one essential part of the design process risks being overlooked: the people we design for.

While AI can help us generate ideas, create prototypes, summarize interviews, and even write production-ready code, it cannot experience frustration, confusion, delight, hesitation, or trust. The human behaviors that drive product success remain outside the reach of even the most sophisticated models. This is why user research is not becoming obsolete—it is becoming increasingly important.

AI is transforming how we work. But it does not eliminate the need to understand users.

As design becomes faster, assumptions spread faster, prototypes multiply faster and features ship faster, meaning that the cost of being wrong increases. Therefore research becomes even more critical.

In the age of AI people might assume that more AI means less research. However more AI implies more research too because speed amplifies both good and bad decisions.

The Workflow is Changing

The design workflow has changed. Before, the classic approach to it started with user and market research → wireframes →prototypes → test → developer handoff (an iterative process).

Today, many teams augment these same activities with AI-powered tools. Researchers use AI to prepare discussion guides, designers use it to explore concepts and generate prototypes, and developers use it to accelerate implementation. While the workflow may look different, the underlying goal remains unchanged: understanding user needs and translating them into valuable solutions.

This evolution isn’t inherently positive or negative, but rather a “natural” evolution given the current times. It’s simply happening.

What AI does exceptionally well?

AI excels at accelerating repetitive work (research planning, desk research, live note taking and transcriptions, multilingual translation, synthesis and analysis…). AI helps us to generate multiple alternatives much quicker, brainstorming multiple ideas in a more organized manner, challenging our assumptions and so on. It also reduces work friction when creating prototypes, drafting documentation, and preparing research scripts. And we, as designers, should embrace these efficiencies.

Consider a usability study involving ten participants. Traditionally, reviewing recordings, extracting notes, identifying patterns, and preparing a report could take days. Today, AI can transcribe conversations, generate summaries, cluster recurring themes, and surface notable quotes within minutes.

This allows researchers to spend less time organizing information and more time interpreting what the findings actually mean.

What AI cannot tell us?

AI cannot reliably tell us: what users actually prioritize, why they behave the way they do, what workarounds they’ve developed, or which problems are worth solving. It lacks context and empathy. Because those insights emerge through observation, interviews, contextual inquiry and usability testing. That is, the most important design questions remain fundamentally human.

AI can identify patterns from existing information, but it cannot observe a participant hesitating before clicking a button, notice uncertainty in their voice during an interview, or understand the organizational constraints influencing their decisions.

For example, users often say they want one thing but behave differently when performing real tasks. This gap between stated behavior and actual behavior has always been one of the most valuable areas of UX research—and one that requires direct observation.

The richest insights often emerge from unexpected moments: a workaround that nobody anticipated, a misunderstanding caused by terminology, or a hidden frustration that never appeared in analytics dashboards. These discoveries come from engaging with users directly, not from generating another prompt.

Faster design increases the cost of assumptions

Before, teams moved slowly, and bad assumptions took time to implement.

Now, AI allows teams to generate dozens of concepts, prototypes, interfaces, and features within hours, which means teams can now scale assumptions at unprecedented speed. If those assumptions aren’t validated, you simply get to the wrong answer faster.

AI does not reduce the need for research; it reduces the excuses for skipping it.

Imagine a product team assumes users want an AI-powered dashboard. With modern tools, they can design, prototype, and build that dashboard in a matter of days.

But if users actually need better reporting rather than more automation, the team may spend significant effort solving the wrong problem.

AI dramatically reduces the cost of creating solutions. It does not reduce the importance of validating whether those solutions address real user needs.

The role of user research is evolving

Research itself is changing. AI can support researchers through discussion guides, recruitment screening questions, or hypothesis generation.

During the research process we work with transcriptions, use note capture tools, and capture observations. After the research process, we cluster insights, prepare summaries, and identify patterns.

But AI should augment –not replace– human interpretation. Researchers still provide context, judgement, empathy, and prioritization.

AI can help transform raw data into organized information. However, information is not the same as insight.

An AI tool may identify that six participants mentioned difficulty completing a task. A researcher's role is to understand why that difficulty exists, whether it reflects a broader behavioral pattern, and what implications it has for product strategy. Insight requires interpretation, judgment, and context—qualities that remain deeply human.

The designer value is becoming clearer

As execution becomes easier, designers contribute through four main areas:

Curiosity

AI can answer questions, but designers must determine which questions are worth asking.

Critical Thinking

AI-generated outputs can appear convincing while still being incorrect, biased, or disconnected from user needs. Designers must evaluate suggestions rather than accept them at face value.

Facilitation

As products become more complex, designers increasingly act as facilitators between users, stakeholders, engineers, and business teams.

Human Understanding

Empathy remains one of the most valuable skills in design. Understanding motivations, fears, goals, and behaviors allows teams to build products that genuinely solve problems.

The future designer may spend less time producing artifacts and more time guiding decisions. As AI automates portions of execution, the designer's role shifts toward understanding complexity, aligning stakeholders, and advocating for users.