Artificial intelligence can help research companies cut costs and accelerate consumer studies, but replacing human-led interviews with synthetic respondents risks leaving businesses with more data and less meaningful understanding of their customers, market research expert Carol Fitzgerald has warned.
Fitzgerald, founder of online research company BuzzBack and a 2026 Insight250 award winner, said organisations must distinguish between using AI to improve research efficiency and relying on technology to replicate the depth, emotional understanding and unexpected discoveries that emerge from conversations with real people.
According to Brandspur Brand News, Fitzgerald made the remarks in an interview published by Research World, ESOMAR’s industry publication, on October 7, 2026, as the research sector continues to explore how artificial intelligence can expand qualitative research without compromising its quality.
She argued that much of the delay associated with traditional qualitative research comes from administrative tasks rather than the interviews themselves. Finding suitable moderators, negotiating contracts, arranging schedules and coordinating stakeholders can consume as much as 20 hours per project, slowing delivery and increasing costs.
Fitzgerald said BuzzBack developed its Studio platform to address these challenges by bringing moderator selection, recruitment, fieldwork and reporting into a more integrated process. The platform offers access to a curated network of more than 50 vetted moderators, with profiles, introductory videos, client reviews and availability information to help organisations select suitable professionals.
The company says Studio can reduce the time required to move from a request for proposal to a final report to 10 days, allowing clients to obtain detailed research more quickly without abandoning expert-led interviews.
Fitzgerald maintained that artificial intelligence is particularly useful for tasks such as analysing transcripts, grouping recurring themes and supporting the synthesis of research findings. Automating these activities, she said, gives researchers more time to interpret evidence, question assumptions and advise businesses on strategic decisions.
However, she cautioned against treating AI-moderated interviews and synthetic respondents as direct equivalents of traditional qualitative research.
Synthetic respondents are generated using existing data, research findings or responses to simulate the views of particular audiences. While Fitzgerald acknowledged that these tools could be useful for preliminary testing, rapid feedback and research involving specialised groups with limited populations, she questioned their suitability for investigations that require genuinely new insights.
Her concern is that simulated participants may reproduce patterns already present in the information used to create them, making it harder to uncover perspectives that challenge a company’s existing beliefs.
Human interviews, by contrast, allow moderators to respond to participants in real time, pursue unexpected comments and explore sensitive issues that may require careful questioning. These interactions can reveal motivations and contradictions that a predetermined sequence of automated questions might miss.
Fitzgerald also warned that pressure to deliver research more cheaply and quickly could encourage businesses to accept methods that do not adequately answer their questions. She stressed that the appropriate approach depends on the purpose of each study, with exploratory research requiring a different level of depth from straightforward product or concept testing.
The debate has implications for businesses seeking to understand consumers as AI tools become more accessible. Faster research may help companies respond to changing preferences and make decisions more efficiently, but the quality of those decisions still depends on whether the evidence reflects the people they intend to understand.
Fitzgerald said the industry’s long-term credibility would depend on being transparent about what AI-powered research can deliver and where human expertise remains essential. She argued that technology should make rigorous qualitative research easier to access, rather than create the impression that simulated conversations can always provide the same understanding as speaking directly with consumers.
