Why User Research still needs a human touch

In a research project's early stages, AI can streamline tasks, giving human researchers more space to shape the strategic direction and drive meaningful change.

Its ability to analyse large datasets to identify trends, gaps, and emerging themes helps shape research priorities. However, the researcher is still critical at this first stage of setting objectives and questions. Understanding user needs and business goals shapes the research strategy, with objectives serving as the guiding star to ensure relevance and impact—something AI cannot replicate.

Once foundational objectives are set, AI can assist in constructing research plans. Acting as a valuable first-drafter, AI enables researchers to generate comprehensive plans using project objectives and background information. This efficiency frees time for strategic refinement rather than starting from scratch. Similarly, AI supports the development of discussion guides and research tasks, using the refined research plan as a brief to ideate core questions, prompts, and task instructions, providing a baseline for researcher iteration.

Perhaps one of AI's most valuable contributions is its role as a "sparring partner" for researchers—particularly beneficial for solo practitioners in smaller organisations. By drafting materials independently and comparing them with AI-generated alternatives, researchers can identify gaps, unexplored angles, or alternative phrasing to refine research materials. This collaborative approach supports and inspires researchers while maintaining efficiency.

In recruitment, AI can streamline the creation of participant screeners based on target audience requirements. Some platforms, like User Interviews, use AI to improve participant matching, while machine learning algorithms assess trustworthiness, ensuring data quality—critical when researching health conditions or specialisms.

Ultimately, human expertise remains essential as a part of research design and recruitment. View AI as an assistant, not a replacement, requiring researchers to evaluate and refine its output. By leveraging AI for drafting and ideation while maintaining human judgment for strategy, we save time without compromising quality.

In the execution phase, AI tools are increasingly able to support user research, particularly in transcription and note-taking. Tools like Looppanel and Otter.ai claim high accuracy in transcribing user interviews, and automatically tag key insights and themes, significantly reducing time spent on these tasks. Session summaries can also be generated, providing updates to stakeholders.

However, while AI streamlines these processes, there are limitations. AI tools often struggle with context and speaker identification, leading to misunderstandings. For this reason, researchers must always double-check AI-generated outputs.

Additionally, AI's inability to interpret user actions or non-verbal cues makes it unsuitable for usability testing or behavioral research. While these tools transcribe dialogue, they fail to capture how users interact with interfaces—an essential aspect of usability studies.

AI in this field is developing quickly. Emerging tools like Versive conduct semi-structured interviews at scale, generating custom follow-up questions. While this offers scalability, it lacks the rapport-building ability necessary for effective interviews. This is especially important in sensitive areas like healthcare, where empathy and human connection create a comfortable environment for participants.

AI created users

Another area of debate is synthetic users—AI-generated profiles simulating user perspectives. While they help generate hypotheses and synthesise broad data, they often provide unrealistic or overly favorable feedback. Similar to AI’s role in interviews, its limitations are clear in niche areas like healthcare, where rare diseases and nuanced patient experiences require real-user insights. Ultimately, ‘UX without real-user research isn’t UX’—synthetic users may supplement, but they should never replace the real thing.

Therefore, while AI is valuable for transcription and note-taking, the human role remains irreplaceable, with empathy, connection, and real-user experiences in mind.

When it comes to analysis and synthesis, AI is commonly used to identify patterns within large datasets, saving significant time when compared to manual synthesis.

It uses intelligent thematic tagging to categorise and organise data into themes and sub-topics directly from transcripts. AI can also group sticky notes in Miro or FigJam, automating affinity mapping. This is a big time saver when synthesising moderated research sessions or open-text survey responses with extensive data to sort through.

However, while AI accelerates initial synthesis, it lacks the ability to interpret meaning beyond surface-level patterns. AI struggles with nuance, irony, or cultural context—factors crucial in user research, particularly in healthcare. Additionally, automated clustering can group insights incorrectly or place many notes into an ‘Other’ category, requiring researchers to refine the outputs.

The power of analysis lies in a researcher’s ability to interpret data holistically, connecting findings to business goals and user needs. AI can help surface patterns and make large datasets manageable, but it cannot determine which insights truly matter. As this phase informs final outputs and recommendations, researchers must review AI-generated analysis to ensure accuracy and relevance.

While AI provides a useful first pass at organising data, human ingenuity is essential for drawing meaningful connections and translating insights into action. Researchers—not AI—bring the critical thinking, context, and creativity needed to make user research truly impactful.

Once research is complete, findings must be communicated in a compelling, accessible way. AI assists with report writing, summarisation, and presentation generation, helping distil complex insights into digestible takeaways. Tools like Looppanel automate executive summaries, while Maze generates AI-driven reports with heatmaps and sentiment analysis. AI can also support artefact creation, such as personas and journey maps, using research data to speed up initial drafts. However, as with all AI outputs, these require careful review to ensure accuracy and remove inconsistencies.

AI chatbots are valuable writing assistants, refining language, adjusting tone, translating UX jargon, and tailoring messaging for different audiences. Communicating findings effectively means simplifying language while ensuring insights resonate with each stakeholder group. AI writing tools like Claude enable quick iteration based on style guides and personas, but researchers must refine outputs through an iterative process—reviewing AI-generated content, making adjustments, and providing rationale until the final output is clear, credible, and fit for purpose. With AI handling much of the iteration, this refinement process is significantly faster.

Beyond reporting, AI improves long-term knowledge management and research socialisation. Research repositories can be difficult for non-researchers to navigate, limiting engagement. AI-powered repositories with smart search functions allow stakeholders to ask questions and receive AI-curated responses highlighting notes, tags, and clips, streamlining access to relevant findings. These tools also assist researchers by automating insight tagging, though human oversight remains essential for accuracy and relevance.

AI’s role at the end of research should not be overlooked. Acting as an assistant, AI accelerates iteration and tailors outputs, allowing researchers to perfect reports and drive stakeholder impact without adding time or cost.

AI is already proving to be a powerful research assistant—automating time-consuming tasks, speeding up synthesis, and making insights more accessible. However, while AI enhances efficiency, it cannot replace the depth of human intuition, strategic thinking, and ethical judgment that UX research demands. 

The most effective use of AI lies in collaboration: letting AI handle the repetitive and scalable aspects while researchers focus on interpretation, creativity, and ensuring insights drive real impact. The future of UX research isn’t AI versus humans—it’s about striking the right balance to maximise both efficiency and quality in the pursuit of better user experiences.