How UX research methods are evolving for AI

As expectations change, so do research methods. Traditional usability testing still plays a role, but it’s no longer sufficient on its own.

Priyanka describes a more nuanced approach to UX research methods tailored for AI:

• Evaluating behavioral signals, not just self-reported feedback
• Tracking emotional responses like hesitation or frustration
• Measuring how trust evolves over time within a session
• Assessing AI outputs for tone, accuracy, and appropriateness

“One approach is behavioral coding of recordings,” she explains. “You go through interactions and identify moments of hesitation, confusion, frustration, and disengagement.”

This approach reflects a broader shift toward emotional UX research, where what users feel becomes just as important as what they do.

Instead of relying solely on surveys or ratings, researchers are analyzing real interactions—watching where users pause, question, or disengage. These micro-moments often reveal more than any post-task questionnaire.

Why consistency and transparency matter more than ever

AI systems don’t just need to be correct—they need to be consistently correct. A single unexpected or incorrect response can erode trust quickly.

Priyanka highlights consistency as a key pillar of AI user experience:

“Inconsistency directly erodes trust,” she says. “You need to evaluate how the AI responds across different users, tasks, and sequences.”

This is especially important in areas like voice AI user experience, where interactions feel more human and expectations are higher. When users are speaking naturally, they expect the system to understand nuance—and respond appropriately every time.

Transparency also plays a major role. Users need to understand what the AI is doing, why it’s doing it, and when it might be wrong. This is where concepts like human in the loop AI become essential—ensuring users remain in control rather than feeling replaced.

The role of accessibility and inclusive design in AI

AI has the potential to make digital experiences more accessible—but only if it’s designed with inclusivity in mind from the start.

Priyanka is a strong advocate for accessibility in AI and inclusive design AI, emphasizing the importance of recruiting diverse participants and testing across a wide range of user needs.

“I start from an inclusive recruit,” she explains. “Diverse participants in how they speak, the language they use, their interaction styles—and then evaluate how the AI responds to each.”

This approach ensures that AI systems don’t just work for a narrow audience but deliver value across different abilities, backgrounds, and contexts.

It also helps uncover edge cases that might otherwise go unnoticed—cases that can significantly impact trust and usability.

From tactical research to strategic impact

One of the biggest challenges researchers face is moving beyond validation and into strategy. In fast-moving product environments, it’s easy to fall into the role of simply testing what’s already been built.

But Priyanka argues that researchers play a critical role in shaping AI product design from the ground up.

“It’s more important than ever for researchers to provide that foundation of your AI innovation,” she says.

She emphasizes the importance of:

• Building strong cross-functional partnerships
• Participating in early-stage product discussions
• Challenging assumptions with user data
• Connecting insights to business outcomes

Interestingly, she also challenges the idea that tactical and strategic work are separate.

“Something as small as changing where your AI presence appears can have a huge strategic impact,” she notes, pointing to how even minor design decisions can influence transparency and trust.

Using AI to enhance research workflows

While researchers are evaluating AI, they’re also increasingly using it to improve their own workflows.

Priyanka describes how tools like Copilot and research analysis platforms are helping streamline tasks like:

• Survey creation
• Thematic analysis
• Insight generation

These tools allow researchers to move faster—but not at the expense of human judgment.

“I think of it as humans partnering with AI,” she says. “The AI can help you generate themes, but you still bring your own judgment.”

This balance is key. AI can accelerate AI research strategies, but it can’t replace the contextual understanding that comes from direct interaction with users.

Building AI fluency as a core skill

As AI becomes central to product development, researchers—and product teams more broadly—need to develop a new kind of literacy.

Priyanka calls this AI fluency: an understanding of how AI systems work, how they fail, and how to evaluate them effectively.

This includes familiarity with concepts like:

• Hallucinations
• Prompt design
• Model evaluation
• Bias and guardrails

“Being AI fluent cannot be an afterthought,” she says. “You want to be able to follow the conversation and ask the right questions.”

This knowledge empowers teams to make better decisions—and to build products that are both innovative and responsible.

The growing importance of product-market fit in AI

Another emerging theme is the importance of AI product market fit. With so many AI-powered features being developed rapidly, not all of them deliver real value.

Priyanka frames this as a critical question for teams:

“Does your AI truly add value to a workflow? Is it transformative? And at minimum, is it trustworthy enough for users to actually want to use it?”

This perspective shifts the focus from novelty to impact. Just because something can be built doesn’t mean it should be.

Teams need to ensure that AI features solve meaningful problems—and do so in a way that users trust and understand.

Human judgment remains the differentiator

AI is changing how products are built, tested, and experienced—but it’s not replacing the need for human insight. If anything, it’s making it more important.

The most effective teams are those that combine the speed and scale of AI with the empathy and judgment of human researchers.

Priyanka puts it simply:

“It’s not about AI replacing what researchers do. It’s about humans partnering with AI to become more efficient and productive.”

That partnership is what enables teams to design experiences that are not only functional, but meaningful.

Designing for trust is designing for the future

As AI continues to evolve, the definition of a great user experience will continue to shift. Usability will always matter—but trust will be the deciding factor.

Designing for trust means:

• Prioritizing transparency over complexity
• Valuing consistency over novelty
• Centering human needs over technical capabilities

It’s a shift that requires new methods, new mindsets, and a deeper commitment to understanding users.

And for researchers, it’s an opportunity to play a more strategic role than ever before.

As Priyanka reflects:

“It's not just about building a great experience anymore—it's about building something users genuinely trust.”

Episode links

• Priyanka on LinkedIn

• Priyanka as an Instagram Content Creator posting about AI, Product, UX: @uxr.pri 

• I'm a senior UX researcher at Microsoft. Here's how I broke into AI without a tech background — and 3 lessons I learned.

• Priyanka speaking at University of Washington

• Priyanka speaking at Grace Hopper Conference

• Priyanka speaking at Women Tech Conference 

• Priyanka delivering a Webinar on Breaking into UX in the Age of AI

• Priyanka on AI Snack Club

• Webinar: Fix the gaps: how to optimize the customer journey for better ROAS and higher ROI

• Guide: Unlocking audience insights: driving smarter decisions across product, marketing, and UX

• Podcast episode: How AI in UX research is augmenting (not replacing) human insight

• Blog post: Know your audience for a better ROI