In today's rapidly evolving tech landscape, AI is revolutionizing how market researchers gather insights and deliver value. We recently sat down with Michael, a veteran researcher with over 30 years of experience who now leads research initiatives at Radius Insights, to get his take on these changes.
Michael’s journey from academic researcher to market research entrepreneur offers a unique window into both where the industry has been and where it's headed next. His perspective on how AI tools enhance—rather than replace—human expertise is particularly valuable for anyone working in or adjacent to market research.
Falling Into Market Research
Like many in the field, Michael never planned to become a market researcher. "I was in graduate school, getting my PhD in experimental psychology with a focus on quantitative methods and cognitive psych," he explains. "I realized academia wasn't for me, so I needed to find a job." This was in the pre-internet days, a time when job hunting involved scouring physical publications.
Michael found a job listing that matched his statistical expertise. "I understood all the statistical methods they described, but had no idea what the field was. It was called customer satisfaction measurement." He landed the job in Indianapolis and began his career.
After building his career across several companies, including a stint at Hewlett Packard, Michael eventually started his own firm, Probit, which he ran for over a decade before being acquired by Radius Insights. "I kinda fell into this field as many people do," he reflects, "not really knowing what it was or what I was getting myself into, but it's been a fun ride."
Building a Business: The Probit Years
Starting a market research firm came with significant challenges. When asked about the biggest hurdles:
- Cash flow management: "Cash flow is always king. At first, it's just trying to gain some traction, get business." Michael knew he had potential clients from his network, but "it's just a matter of proving yourself, and getting those first few wins."
- Workload balance: "As you grow... how do I manage all of these projects, all of this work? Working literally eighteen hours a day, six to seven days a week for many years" to get the business into rhythm.
- Team building: "Finding and hiring the right staff—I was lucky because I had a really good network of people I'd worked with. All the people I ended up hiring were people I'd worked with in the past, and knew well."
Probit operated as a virtual company long before remote work was the norm. "We were a virtual company, never had an office," Michael recalls. "This was well before the Zoom days." Managing the workload was critical, but the team found a way to work together seamlessly.
Their competitive edge came from three key areas:
- Statistical expertise and data science capabilities
- Exceptional customer service and quick turnarounds
- Deep understanding of technology
"We were willing to go above and beyond for our clients," Michael notes. "Working long hours, turning around data more quickly, providing higher levels of customer service than typical large firms."
How AI is Transforming Market Research
Michael sees AI reshaping market research in three fundamental ways:
1. Boosting Efficiency
"AI is going to allow us to do things much more quickly than we've been able to do before." This includes speeding up analysis, automating parts of report generation, and streamlining workflows, freeing up researchers' time for higher-level tasks.
2. Elevating Insights
"Rather than spending time in minutiae, we'll be able to really think about the findings and provide richer insights to our clients." Instead of just reporting data points, researchers can focus on interpretation, strategic implications, and telling a compelling story with the data.
3. Deepening Respondent Engagement
"Through intelligent probing on surveys or AI-based conversations, we're able to engage with respondents in a different way than with just a static survey." Michael notes that standard open-ended survey responses are often "pretty flat," lacking depth. AI helps overcome this by asking relevant follow-up questions in real-time.
These aren't just theoretical benefits. Michael’s team has seen measurable improvements with their "intelligent probing" approach.
"With AI-based follow-up probing, we find that we get much richer insights. We identify about 20-25% more themes. It's really helping us understand the 'why' behind responses."
The Human-AI Partnership
Michael has been evaluating tools for coding open-ended responses for over a decade. "I've evaluated I don't know how many different tools over that span of time. All were machine learning, more machine learning-based tools, not necessarily AI, you know, neural network-based tools. And over that testing period, I was never terribly impressed with the quality of the output from a machine-based coding tool," he mentions.
The game changed with modern AI. "But that really has started to change with the LLMs [Large Language Models]. What we're seeing is that the large language models really provide much richer, better quality of coding in comparison to what we've seen." Radius has now developed its own in-house coding tool leveraging these advancements.
Despite his enthusiasm for AI tools, Michael emphasizes one crucial principle:
"You have to have the human in the loop."
His team's approach to analyzing open-ended responses illustrates this partnership:
1. Run unstructured responses through AI to identify potential themes/codes.
2. Have experienced researchers evaluate, combine, and refine those AI-suggested codes.
3. Use AI again to apply the refined coding framework consistently across the dataset.
"This hybrid approach works much better than just simply having AI go off and do the coding," he explains. "AI is a good tool, but you need experienced researchers in that process to guide it and validate its output." The same principle applies to qualitative research, where AI tools can assist with initial analysis, theme identification, and finding relevant quotes, acting as "a second person researchers can bounce ideas off of."
Navigating Controversial Territory: Synthetic Data
When it comes to synthetic respondents (AI-generated survey participants), Michael acknowledges it's "by far the most controversial topic in market research in relation to AI." Many researchers are unfamiliar with the concept or wary of its implications.
He defines synthetic respondents as "digital twins modeled after real respondents" and outlines a responsible approach:
"First, conduct primary research with real people to understand their attitudes, behaviors, and demographics. Then feed those metrics into AI to have it assume the persona of each individual."
This approach can be valuable, "especially when you're trying to reach low incidence populations, or you're trying to do a lot of iterative testing," but Michael stresses the ethical considerations:
- Transparency is key: "Be transparent with your clients, the industry, and respondents."
- Ethical grounding: "Be ethical in your approach." This means not misrepresenting synthetic data as human data and understanding its limitations.
- Build on real data: Synthetic respondents should ideally be based on and validated against real human input, not created in a vacuum.
Ensuring Data Quality in the Age of AI
Data quality has always been challenging in market research, especially with the rise of bots and fraudulent respondents, particularly in B2B studies where incentives are higher. "It's a huge issue," Michael confirms. AI offers new tools to combat this:
- AI-powered screening: "We're embedding open-end questions in screeners and using AI to evaluate response quality, relevance, meaningfulness, and whether it was likely written by a human or AI." Michael mentioned an upcoming pilot with a toy company to see if AI can distinguish real children’s responses from bots or parents filling out surveys.
- Back-end verification: "We use a combination of AI and machine learning to score respondents based on their entire survey response pattern and identify potentially problematic ones."
- Human verification: "We still go record by record among flagged respondents to see if they pass the smell test. Even if they made it through the screener, subsequent answers might reveal they aren't a qualified respondent."
This multi-layered approach, combining AI detection with experienced researcher judgment, is crucial for ensuring research is based on legitimate, quality responses.
Bridging Research and Technology
Michael’s background in both research and programming (he minored in computer science) gives him a unique advantage when working with technical teams.
"I'm able to think a bit like a programmer... Not necessarily a good programmer, but a programmer," he jokes. "I'm able to bridge the gap between programming and market research." This understanding facilitates collaboration and troubleshooting.
"When we find that a tool isn't working quite right, we can problem-solve and figure out how to enhance its accuracy," he explains, mentioning adjustments to AI parameters like 'temperature' to fine-tune performance. Having a development team that also understands market research is a significant asset.
Advice for Future Market Researchers
For those entering the field today, Michael offers clear guidance on staying relevant in an AI-powered future: "As AI becomes more prevalent, our value as researchers will be in synthesizing information—say, across different studies, or synthesizing information within a single study—and providing recommendations beyond what's written on the page."
The key is developing skills that complement rather than compete with AI:
- Synthesis skills: Connecting insights across multiple data sources and studies.
- Strategic thinking: Moving beyond summarization to actionable recommendations and implications.
- Consultative value: Providing the contextual understanding, business acumen, and storytelling that AI currently lacks.
"AI will do a great job summarizing research, but the 'so what' and the 'why'—it may not understand as well. We need to provide value through consultation and strategic thinking."
The Bottom Line
AI isn't replacing market researchers—it's transforming how they work and the value they deliver. By automating routine tasks and enhancing data collection and analysis, AI frees researchers to focus on deeper insights, strategic recommendations, and building consultative relationships with clients.
The most successful researchers will be those who embrace AI as a powerful tool while developing uniquely human skills of synthesis, critical thinking, contextual understanding, and strategic application.
As Michael’s journey shows, the market research field continues to evolve. Those who can adapt to new technologies while maintaining focus on fundamental research principles and delivering true consultative value will thrive in this new era.
