Watch this WaysConf session, then continue with related talks or explore the current programme.
In this panel discussion from WaysConf 2025, researchers Jakub Serek, Weronika Denisiewicz, and Małgorzata Sapeta explore the evolving role of tools in UX and CX research. They define what constitutes a research tool—ranging from physical notebooks and communication apps like WhatsApp to advanced AI platforms like NotebookLM—and contrast tools with research methodologies. The speakers share their personal tool stacks, emphasizing that tools should remain secondary to research goals and researcher competence. They discuss the balance between specialized and general-purpose tools, advocating for a minimalist, validation-first approach before investing in expensive software. Finally, the panel dives deep into the integration of generative AI, examining its utility as a critical sparring partner, an analytical aid for large datasets, and a report-writing assistant, while highlighting its limitations in capturing human contradictions, emotional nuances, and contextual subtleties.
Tools are the execution mechanisms, whereas methods represent the strategic approach. A tool should never dictate the research design but should serve as an aid to optimize resources and manage data.
Watch from 18:04Before adopting expensive, specialized research tools, teams should validate their processes using simple, low-cost alternatives like spreadsheets. Only scale to complex platforms when manual data management becomes a bottleneck.
Watch from 27:46Researchers do not need to limit themselves to dedicated research software. Everyday communication tools like WhatsApp can be creatively adapted for methodologies like ethnographic diaries, provided they meet security and compliance standards.
Watch from 19:27Generative AI is highly effective when prompted to act as a critical reviewer rather than a passive assistant. It helps identify logical gaps in research plans and generates diverse scenario drafts for testing.
Watch from 50:28While synthetic personas can help generate initial hypotheses, they often produce overly logical and sanitized profiles. They fail to capture the inherent contradictions, emotional nuances, and illogical behaviors of real human users.
Watch from 56:44These concise notes were edited from automatic captions and checked against the talk structure. They are not a verbatim transcript.
Małgorzata and Weronika explain that a research tool is anything that assists in planning, executing, or summarizing a study. This definition extends far beyond specialized software to include physical items like notebooks, workshop materials like Dixit cards, and even frameworks like customer journey maps. They emphasize that tools are secondary to the research itself; they should never dictate how a study is run but should instead act as a supportive addition to the researcher's core knowledge and expertise.
Watch from 11:53Małgorzata shares a practical example of using WhatsApp to conduct ethnographic diaries. Although the messaging app was never designed for user research, it served as an excellent, low-barrier tool for participants to submit screenshots and comments. While this creative approach made participation incredibly easy for the users, it did require significant manual effort from the research team to transfer and organize the data afterward, illustrating the trade-offs of adapting non-standard tools.
Watch from 19:14Weronika advocates for a lean, validation-first approach when selecting research tools. In a previous role, instead of immediately requesting a budget for expensive software to track user feedback, she began by manually logging and tagging feature requests in a basic spreadsheet. Once she proved that this data actively helped product managers make better decisions, she had the leverage and justification needed to invest in more sophisticated, scalable tools.
Watch from 27:46Małgorzata warns against using AI to analyze qualitative data if the researcher did not personally conduct the interviews. She describes a case where a colleague tried to analyze another researcher's transcripts using AI, resulting in highly inaccurate outputs. The issue was resolved only when the colleague realized they had to query the AI using the specific vocabulary and conversational style of the original interviewer, highlighting how AI struggles with personal nuances.
Watch from 44:00Jakub explains his preference for using generative AI as a critical sparring partner rather than a source of praise. While default AI models tend to compliment user drafts, Jakub intentionally prompts the AI to be highly critical of his research plans. This approach helps him identify blind spots and logical gaps in his methodology before launching a study, demonstrating how AI can actively improve research quality rather than just saving time.
Watch from 50:28The panel addresses the concept of synthetic personas generated by AI. While they agree these models might offer a decent starting point for formulating initial hypotheses, they remain highly skeptical of using them to replace real users. Real humans are inherently contradictory, emotional, and often illogical in their decision-making. Because AI models generate highly polished, logical, and sanitized responses, they fail to capture the messy realities of human behavior.
Watch from 55:23
