AI and UX Research: how I'm integrating AI into my work (and the resources that have helped me)

Resources, critical voices, and open questions from practice, not from hype.

Image: The Pony Tramway, from the book Abroad (1882) by Thomas Crane and Ellen Houghton. Public domain via Project Gutenberg.

---

How I got here

I started using ChatGPT two or three years ago, for fairly basic things: editing texts, writing emails (not systematically), answering questions almost as if it were Google. Then I discovered GPTs and started creating them for specific documents or projects, but back then they didn’t integrate well — they couldn’t read across conversations, and it was exhausting to repeat the same context over and over. Still, I used it.

At the start of 2025 I found myself out of work. Instead of jumping straight into job hunting, I decided to explore. I started a YouTube channel with my daughter in mind and made content with AI: videos, music, voices (mine and generated ones), characters, stories, tales. It was an intense period of learning. I opened another channel in another country to test reach. I kept exploring tools. Around that time I seriously considered not returning to UX Research.

I had a mini existential crisis: I didn’t know whether to go back to UXR, look for work as a psychologist, or look into something completely new. Because something my first career change taught me — from clinical psychology to UX Research — is that we’re far more flexible than we think. Skills transfer from one place to another. We can start over. I asked my coach for support (someone I turn to at certain moments, not on an ongoing basis), and she asked me something like: “but what is it that fills you? What would you do even if no one paid you?” The answer was clear: sharing what I know, even if it’s just a little, with others. Telling everything I’ve learned. That motivated me enormously.

Shortly after, I landed an ongoing project with a client. I’m still working with that client, and others have appeared more occasionally. Things have never gone this well for me. It seems that connecting with what truly moves you helps — or the planets aligned. Probably a bit of both.

I returned to UXR, and I returned with AI. I kept trying tools — ChatGPT and Gemini, both paid — and at some point I switched to Anthropic. My life lit up (genuinely). I unsubscribed from ChatGPT. I went from talking to a forgetful junior to talking to a peer colleague — a somewhat scattered one, but one who gets it. And I started using it just when I picked up user research again, content creation (because I also work as a content SEO manager), and exploring what Gemini GEMs and Claude Projects could do.

Today I work almost every day on something related to this site — whether it’s organization, brand consistency, or a blog post — with Claude and Claude Code, alongside Gemini. I haven’t been able to use Cowork because my machine doesn’t support it, but I do use the chat and Claude Code. That’s my current working setup, and it’s still a work in progress.

This post comes from that: months of reading, attending talks, downloading guides, testing tools on real projects. I have no certainties about how AI “should” be used in UX Research. What I have is a map of resources that has served me, and a group of voices that I find more honest than average on the subject.

If you’re looking for a general guide to resources for learning UX Research, I wrote a complete guide here. This post zooms in on a specific topic: how AI is intersecting with research work, and what I’m learning along the way.

---

What I’m seeing in the conversation

There’s a group of practitioners who are using AI in their research work and sharing what they discover — with nuance, not hype. I follow them on LinkedIn, read their newsletters, attend their talks when I can. What strikes me is that many are reaching similar conclusions without having coordinated.

For context: according to a recent essay published in the Journal of User Experience (Sokolov, 2026), AI tool adoption among researchers jumped from 20% in 2023 to 50% in 2024. The projection for 2026 is 80%. This isn’t a niche — it’s something happening across the entire discipline. But the conversation about how to use these tools well, in my experience, is still in its early stages.

Using AI forces you to make explicit what was previously tacit

That, in my experience, is the thread connecting almost everyone.

JB Booth, UX Research Leader focused on conversational agents, wrote in his newsletter Talking Systems about how he automated RITE study reporting with Claude. After eight rounds of editing, he realized the corrections were more valuable than the report itself: each edit surfaced an unwritten rule about how he interprets findings, weighs evidence, and communicates results. His phrase that kept echoing for me: “AI doesn’t replace research judgment. It forces you to write it down.”

Nikki Anderson, founder of The User Research Strategist, has built over 52 SKILL.md files for research in Claude (essentially structured instructions that tell Claude how to execute specific research tasks). What she didn’t expect was that the process of writing them made her a better researcher: to create a skill around, say, insight writing, she had to articulate exactly what makes a good insight different from an observation, what criteria she uses to decide whether something is worth including in a report. Things she’d been doing by instinct for years, which she’d never had to put into words. Nikki is a reference point for me — I’ve been following her Substack for a while — and her evolution toward AI applied to research is one of the most honest I’ve seen.

Dr. Susanne Friese, qualitative researcher and founder of QInsights, ran a revealing test with NotebookLM. She gave it four qualitative interviews about friendship formation and asked about participants’ first impressions. NotebookLM jumped straight to thematic synthesis and interpretation — conclusions, not starting points. When she slightly changed the prompt, not just the words changed: the themes changed, the interpretation changed, and the passages the system selected as evidence changed. Worse yet: a case that didn’t fit the question (a participant whose friendship formed through proximity, not through a defining moment) was simply omitted. As Susanne says: “a response that quietly discards the inconvenient case is not grounded analysis — it is selective retrieval disguised as a finding.” For anyone doing qualitative research, that should matter.

Ruby Kuo, UX Director with over 20 years of experience, built an AI agent for UX audits. Within its defined scope, the agent performed well — consistent, structured, reliable. But when the context became ambiguous, it pressed forward with confidence, without signaling that it was on uncertain ground. Her reflection: “The real risk isn’t AI replacing researchers. It’s teams using AI to get to the wrong answer faster.” That deserves attention.

Caitlin Sullivan, former Head of UXR at Spotify Business, insists on something that in my experience almost no one teaches: verification. Not as a step added at the end, but as something built into the workflow from the beginning. The difference, according to her, between fast analysis and analysis you’d put your name on in front of stakeholders. Caitlin has courses on Maven and free workshops where she goes deeper into this (mentioned below).

Constantine P., Senior UX Researcher at Uber, is publishing a book this summer: “AI-Powered UX Research: How to Run Research at the Speed Your Team Actually Needs.” His thesis strikes me as powerful: AI didn’t create the gap between research and product decisions — it exposed it. By compressing the development cycle, it made it impossible to ignore a gap that research had been quietly tolerating for years. That reframes the conversation: the problem isn’t the tool, it’s the structure.

---

Resources for learning

These are the resources I’ve found and that, in my experience, have more substance than average. I organized them by type so you can go straight to what’s useful for you.

Guides and downloadable material

• The Explosion of AI-Powered Research Tools Is Reshaping UX Practice — Jeffrey L. Sokolov, PhD. Essay published in May 2026 in the Journal of User Experience (UXPA). It’s probably the most rigorous source on this list: it examines three interrelated changes (researchers as workflow orchestrators, insights fragmented across tools, and the role of AI in shaping interpretation). Includes concrete data: AI tool adoption among researchers went from 20% in 2023 to 50% in 2024, with a projection of 80% in 2026. If you’re going to read one thing from this list, make it this.

• Benefits and Risks of AI in User Research Analysis — Condens. A framework that presents benefits and risks as two sides of the same coin: speed vs. precision, scalability vs. verification burden, accessibility vs. the risk of missing hallucinations. Useful for having an honest conversation about the topic with your team.

• How to AI UXR — The ResearchOps Review. A maturity map organized into three levels (Crawl, Walk, Run) showing how research professionals are integrating AI into method selection and scoping. From generating templates to building agents that automatically surface previous research. Concrete and practical.

Free courses and workshops

• Caitlin Sullivan on Maven x Lenny — Two on-demand, free workshops: “Claude Code for PMs: Data to Decisions Workflows” and “Synthetic Users for Product Discovery: What People Miss.” Although they say “for PMs,” the content is directly relevant for researchers.

• Pascal Raabe: AI for UX Research — Udemy. Can be audited. Covers the complete workflow: planning, recruitment, interviews, analysis, synthesis, and reporting. Updated for 2026 with worksheets and prompt templates. What I like: it teaches a quality loop (Draft, Critique, Verify, Document) instead of just “put this into ChatGPT.”

• UX Researchers Guild AI Club — Periodic sessions with guests like Natalie Golub (Coinbase, former Airbnb) on using AI to transform static reports into interactive experiences. Sessions are recorded and sent to registered attendees.

• Great Question: Synthetic Users… Should You or Shouldn’t You? — On-demand webinar on synthetic users. A topic that generates debate and where it’s worth hearing arguments before forming an opinion.

Videos and talks

• — LA UXR Meetup, with Khalil Sullivan. I attended this one and it was really good. Khalil presented a 4-step framework for creating AI-assisted workflows (define objectives, map the process, create your prompt engineering framework, and choose model + data), showed the COSTAR framework for structuring prompts, and shared a real case study of rapid evaluation with A/B testing of conversational models. What I take away: it’s one of the few resources I’ve seen where someone shows the complete process — not just “use AI to analyze,” but how to get from goal to output, including what to document when it works and when it doesn’t. He was also honest about limitations: “Deep Research is prone to hallucinations” and “Check the references” were among his take-aways.

• UN Behavioural Science Week 2026 — 12 free videos from the 8th edition, focused on the intersection of AI and behavioral sciences. This isn’t from the pure UX ecosystem, and that’s exactly why I find it valuable: if you come from psychology or social sciences, this kind of content connects your training with what’s happening in AI in a way that “10 AI tools for UX” articles never manage.

• — YouTube playlist with talks on AI adoption in practice.

• Condens: UX Research Meetup and Event Guide — If you want to find more research meetups and events (not just AI-focused), this guide is a good starting point.

Newsletters and voices to follow

If you want to stay current with critical perspective (not hype), these are the voices I follow:

• Nikki Anderson — The User Research Strategist (Substack). Publishes weekly on research, impact, and AI. 15,000+ subscribers.
• JB Booth — Talking Systems (Substack). On conversational agents and evaluative research.
• Caitlin Sullivan — On LinkedIn and Maven. AI for customer insights, without hype.
• Dr. Susanne Friese — On LinkedIn. Qualitative research and AI, with a critical perspective.

---

How I stay up to date

I don’t have a closed system. I have a process under construction, and I think it’s worth sharing because sometimes the question isn’t “which tool do I use” but “how do I keep from getting lost in everything out there.”

Here’s how it works for me:

I follow people on LinkedIn who discuss the topic with critical perspective. I don’t follow accounts that promise “AI is going to change everything” — I follow practitioners who are testing things and reporting what works and what doesn’t. Most of the resources I curated in this post come from there.

I download PDFs and guides when they appear, and read them carefully. I don’t collect just to collect (well, sometimes I do), but the Condens and ResearchOps Review guides, for example, I read completely and they helped me understand the landscape.

I sign up for free workshops and watch the recordings. The Maven x Lenny series, the UX Researchers Guild AI Club, Great Question webinars. I can’t attend everything live, but recordings are a resource we underestimate.

I watch videos that aren’t just about UXR. Channels like Peter Yang (product), the Claude channel (to understand the tool), Google Workspace (to see how AI integrates into everyday workflows), and Google Career Certificates when they cover AI topics. Seeing how other disciplines use AI gives me perspective I can’t find inside the UX ecosystem.

I attend meetups when the topic is relevant. The LA UXR one on AI workflows was one of the best I’ve seen this year.

I use it actively in my work. Gemini GEM for qualitative corpus analysis. Claude as a methodological and editorial companion — with Projects configured for my consultancy. Claude Code to implement changes on the site. This isn’t theory: it’s daily practice, errors included.

I look outside the UX ecosystem. UN Behavioural Science Week, applied psychology and sociology content, service design. AI is crossing many disciplines, and often the most useful learnings come from outside.

I’m still building my way of working with all of this. I don’t have a closed workflow or a definitive opinion on which tools are “the best.” What I have is curiosity, practice, and a set of sources I keep updating.

---

A closing (without really closing)

If anything is clear to me after these months, it’s that the conversation about AI and UX Research is just taking shape. There’s plenty of content about tools (in the Spanish-language SERP, most articles are tool lists with the same enthusiastic tone), but there’s less content about the deeper questions: what happens to the researcher’s judgment when AI generates convincing outputs, how do we verify what AI returns to us, what tacit knowledge are we discovering we had only because now we have to write it out for a machine.

Those are the questions that interest me. And from what I’ve seen, the answers are coming more from practitioners who experiment and share than from articles that promise “AI is going to revolutionize UX Research.”

If there’s something I didn’t mention that should be here, or if you have a different experience with any of these tools, I’d like to know. Write to me on LinkedIn or subscribe to the newsletter — corrections are always welcome.