We surveyed 1,548 U.S. adults about how they use AI for health decisions, with oversamples of Hispanic, Black, and Asian consumers. Forty-three percent have asked an AI chatbot a health question in the past year. Among Hispanic consumers it's 57%. Among Asian consumers, 54%.
Doctors are already feeling this. Sixty percent of people who've used AI for a health question said it sent them to an appointment they wouldn't have booked otherwise. Sixty-five percent said it gave them a specific drug or treatment name to bring up once they were in the room.
That second number surprised us more than the adoption figure did. Forty-three percent using a chatbot sounds like a curiosity stat but patients showing up already able to name a treatment changes the actual conversation happening in the exam room.
Download the full report here.
Trust hasn't caught up to usage, though. Forty percent of people trust AI chatbots for health information. Trust in Doctors sits at 81%, pharmacists at 70%.
We expected that gap to be about quality. It isn't, at least not according to the people who've actually used these tools: 85% or more of AI health users, across every group we measured, rate the information they got as accurate. So the 40% trust number isn't a verdict on whether AI gets things right. It's a judgement on who's saying it.
We tested that directly. Would you trust an AI health tool built by a big tech company? Thirty-four percent said yes. Would you trust one provided by your own doctor or hospital? Sixty percent said yes.
A 26-point swing on branding alone is a bigger lever than most product features a health system could build this year. If you're a hospital or a pharma brand hesitant to touch AI because it feels like Silicon Valley's territory, this data says the opposite: your name is worth more here than theirs is.
Hispanic and Asian respondents lead nearly every measure in this study: usage, trust, optimism about where AI is headed, interest in new features. Fifty-six percent of Hispanic consumers are optimistic about AI's future, against 47% for the total market. Fifty-seven percent of Asian consumers trust AI chatbots for health info, against 36% of White consumers.
Some of that comes down to language. Fifty-three percent of Hispanic consumers have used AI to get health information in Spanish or another preferred language, and 38% said it worked well. Forty-one percent of Asian consumers have used AI in a heritage language, and 47% said it gave them information they couldn't easily find elsewhere.
We want to flag something in that number, “Worked well” and “worked” aren't the same claim, and we're relying on self-report here, not a clinical accuracy check. Still, for a household that's been stuck translating a pamphlet by hand or waiting weeks for a bilingual appointment slot, “worked well enough to help” is a real result, and it's one users are reporting without much prompting from us. A brand that shows up early with genuine language support in this space is solving a problem that's been difficult to address for a long time.
Black consumers show lower AI usage for health questions (35%, against 43% market-wide), but the reason isn't distrust of the technology. The top answer non-users gave was “I prefer a real doctor.” The second was simply not using AI much at all. Black non-users were also more than twice as likely as Asian or White non-users to say they didn't know AI could help with this kind of question in the first place.
Women lag men on almost everything we asked: usage, trust, comfort, optimism. Only 23% of women who've used AI for health questions have told their doctor about it, compared with 37% of men. Whatever's driving that gap, it isn't education or income, since we controlled for those. It looks more like confidence, and confidence gaps close differently than awareness gaps do.
Boomers are the strangest case in the dataset. Once they try AI for a health question, 90% rate the information as accurate, the highest of any generation. But general trust in AI chatbots among Boomers sits at 27%, the lowest of any generation. They rate it best and trust it least.
One more thing worth noting on the doctor side of this: trust in providers themselves is high and fairly even across the market, with one exception. Black consumers report comfort discussing health concerns with a provider at 82%, ten points below every other group we measured. That gap predates AI and probably says more about the health system than about any chatbot.
AI will continue to shape how consumers find health information and prepare for medical decisions. Thirty-nine percent expect their use of AI for health to increase. The figure reaches 51% among Asian consumers and 44% among Hispanic consumers.
Healthcare organizations should begin with several priorities:
Build AI tools around established relationships with doctors, health systems and trusted health brands. Develop practical features tied to clear consumer needs. Invest in multilingual and culturally informed experiences. Explain privacy protections and information sources. Give consumers clear guidance about the appropriate role of AI in health decisions.
The adoption data shows that consumers have already incorporated AI into their healthcare routines. Their questions, expectations and behaviors will continue to evolve as the technology becomes more familiar.
Healthcare brands now have an opportunity to shape that experience with accuracy, transparency and cultural relevance.
Download the full report here.
This analysis is drawn from ThinkNow's national study of 1,548 U.S. adults, fielded with oversamples of Hispanic, Black, and Asian consumers. Full topline data and methodology available on request.
As artificial intelligence continues to evolve, the conversation is shifting from what AI can generate to what it can do.
In the latest episode of The New Mainstream Podcast, Michael Nevski joins Mario Carrasco to explore the next phase of AI: agentic systems and their implications for consumer behavior, payments, and trust.
Michael Nevski, Director of Global Insights at Visa, brings a unique perspective at the intersection of data, economics, and real-world consumer decision-making. Recognized as one of the most influential professionals in the insights industry, he shares how emerging technologies are reshaping how we understand and interact with consumers.
While generative AI has transformed content creation and automation, agentic AI introduces a new paradigm. These systems don’t just assist, they act.
From making purchases to managing financial decisions, AI agents have the potential to operate on behalf of consumers. This shift raises important questions:
In industries like finance and payments, the implications are even more significant.
As AI begins to participate in transactions, trust becomes a central pillar. Consumers are not just evaluating brands anymore; they are evaluating systems.
This evolution challenges companies to rethink:
For brands, this is not just a technological shift. It’s a behavioral shift.
Understanding how consumers feel about AI acting on their behalf will be critical for future growth.
As Michael highlights, the opportunity lies in translating complex data into actionable insights that help organizations navigate uncertainty, anticipate behavior, and build trust in an AI-driven economy.
The brands that succeed will be those that understand not only the technology, but the human response to it.
Listen to the full episode of The New Mainstream Podcast and explore how agentic AI is shaping the future of consumer behavior.
At ThinkNow, we believe that understanding people starts with listening and getting beyond data points. By integrating artificial Intelligence (AI) into our online panels, we’re transforming how we capture and analyze open-ended responses in market research.
For years, open-text analysis was a manual, costly, and limited process. Today, AI enables us to process qualitative insights with unprecedented speed and precision, optimizing every stage of the research cycle. With these technologies, we don’t just analyze words; we interpret emotions, tone, and context, uncovering the authentic voice of the consumer that traditional methods often miss.
One of the most significant innovations is the ability to collect responses in audio or video format within the panel. This approach allows participants to express themselves more naturally, adding nuances that written text cannot capture. AI transforms these recordings into structured, automatically coded information, available in real time to analysis teams.
Moreover, machine-learning algorithms can assess the coherence and authenticity of responses, enhancing panel quality and reducing human bias. This results in more reliable, representative insights, especially in multicultural studies where expression and context are key to accurate interpretation.
This convergence of AI and online panels ushers in a new era in research, one where the boundaries between quantitative and qualitative blur, giving way to a faster, smarter, and more human ecosystem of insights.
ThinkNow is also expanding these innovations through synthetic sample, an advanced approach that broadens the reach and representativeness of studies without compromising methodological integrity.
If you’d like to learn more about how AI, online panels, and synthetic sampling are revolutionizing research, click here.
Artificial intelligence (AI) is rapidly reshaping society, but with its transformative power comes pressing ethical, cultural, and social questions. The conversation around AI often centers on new capabilities, but equally important are the implications for equity, transparency, and human values.
A key concern is the concentration of AI development in a handful of industries, particularly technology and finance, which risks creating tools that benefit only a narrow segment of society. When innovation prioritizes speed and competition, the so-called “AI race” can result in systems being released prematurely, riddled with bias, or inaccessible to much of the global population.
Language representation in AI models is another critical issue. Many large language models are predominantly trained in English, resulting in the underrepresentation of other languages and cultural perspectives. This imbalance not only limits accessibility but also reduces the quality of AI outputs. Advocates stress that LLMs trained on multicultural data lead to better, more representative systems, ones capable of reflecting the world’s diversity rather than reinforcing existing biases and stereotypes.
Still, the potential for AI to drive positive impact is significant. From creating accessible tools for immigrants navigating new systems to providing voice-based digital companions for older adults, socially conscious applications of AI can foster inclusion and improve quality of life.
On this episode of The New Mainstream podcast, Norman Valdez, CEO of BrainTrainr, discusses the urgency of developing responsible AI and highlights both the dangers of exclusion and the opportunities for technology to serve as a force for good.
We're halfway through 2025 and one thing is undeniable: AI is no longer on the horizon, it is in the room. For the market research industry, this has come faster than most expected. What felt like an existential threat just a year ago is now transforming how researchers approach everything from segmentation to recruitment to data analysis.
But as AI becomes embedded in our workflows, a critical question arises. Are the datasets powering these models truly inclusive? Do they reflect the diverse populations researchers aim to understand, or are they building the next generation of tools on top of the same old blind spots?
Market research has long struggled with inclusivity. Reaching Spanish-dominant Latinos, Gen Z respondents and even male participants has always been difficult. Despite decades of effort, many of these groups continue to be underrepresented in online panels and large-scale studies.
Now, imagine deploying AI on top of these incomplete datasets. Instead of closing representation gaps, AI trained on biased data risks amplifying them at scale. Biases that were once isolated can now be baked into algorithms and amplified across the entire research ecosystem, undermining the potential of AI to drive more inclusive insights.
When AI began gaining traction in the industry, initial skepticism emerged among some researchers, particularly regarding the use of synthetic data and AI-powered moderators. These tools seemed impersonal, disconnected from the human insights that drive understanding and trust among respondents.
Yet, over time, AI has proven itself capable of complementing, rather than replacing, researchers’ work. Instead of diluting what makes insights meaningful, AI can expand them by enabling researchers to finally address representation issues that more conventional methods have never been able to. This shift has prompted a more intentional approach to innovation. If synthetic data is going to shape the future of insights, it must be inclusive by design, representing the full diversity of the populations it aims to model.
The market research industry is uniquely positioned to lead in this space. While many tech companies face lawsuits for training AI on copyrighted or illegally scraped data, researchers have operated under strict privacy laws like GDPR and CCPA for decades. Upholding consent, data stewardship and adherence to ethical standards has been the norm.
Our datasets are not only large, but they are also permission-based and carefully vetted. This makes them ideal for training AI models that need to mirror real-world diversity.
But it is not enough to have access to data. The same rigor applied when building representative samples must be applied to training AI models. This means proactively identifying gaps, asking who is missing from the data and taking measurable steps to responsibly include them.
This brings us to the future of multicultural segmentation. Relying solely on broad demographic categories or historical internal datasets is no longer sufficient. Today’s consumers are multidimensional, and AI gives us the tools to see them more clearly.
To generate synthetic data that accurately reflects multicultural audiences, it is essential to incorporate information from historically underrepresented communities. This requires collaboration between technologists and cultural experts, as well as a commitment to designing systems that accurately reflect the reality of diverse identities.
For researchers generating synthetic datasets, combining privacy-compliant methods with culturally rich data points, powered by AI, helps ensure that communities often left out of the conversation are fully represented moving forward.
AI is not a passing trend. It is here to stay, and it is reshaping how we segment audiences, recruit respondents and activate insights. However, AI’s success depends on the quality and inclusiveness of the data behind it, and the researchers guiding its application.
For market research professionals, this is a challenge worth embracing. With deep expertise, ethical frameworks and a foundation in representative sampling, the industry is uniquely positioned to ensure that AI serves all communities, not just the most accessible ones.
The future of multicultural segmentation will belong to those who successfully integrate innovation and intention because the question is no longer whether to adopt AI, but how to use it in a way that advances representation.
Those investing in synthetic data and inclusive segmentation strategies play a crucial role in achieving this, and those seeking better representation in data must continue to demand it.
This blog post was originally published on Quirk's Media.
For decades, the foundation of market research rested on one powerful tool: the survey. It was the standard way to understand consumers, what they like, want, and feel. Researchers spent years mastering the art of crafting questions, selecting the right sample, and interpreting the answers. And for a long time, that worked well.
But over the last few years, something fundamental has changed.
As the digital world expanded, so did the ways consumers interact with brands. People now browse online stores, leave reviews, post on social media, click on ads, abandon carts, binge-watch videos, and scroll through countless pieces of content. Each of these actions generates a trail of data. These behavioral breadcrumbs reveal more than a simple survey ever could.
But a new era of predictive market research is emerging, one that relies less on what consumers say and more on what their behavior reveals. With the help of predictive analytics, researchers are not just looking at current trends, they’re forecasting future ones.
The shift is happening for good reason. In today’s hyper-competitive, always-on business environment, companies need faster, deeper, and more accurate insights to make decisions. Waiting days or weeks for survey responses isn’t always practical, especially when product launches, ad campaigns, and market shifts happen at the speed of social media. Predictive insights, powered by machine learning and advanced analytics, are giving businesses the edge they need by offering a more dynamic and forward-looking understanding of consumer behavior.
This is especially relevant for industries where consumer expectations shift quickly, like retail, consumer tech, travel, and even healthcare. Imagine being able to predict what your customers are likely to buy next month, which messages will resonate best, or which audience segments are most likely to churn. That’s not science fiction. It’s becoming the reality for modern market research.
The tools driving this shift are growing more advanced every day. Artificial intelligence (AI) can now comb through huge datasets, like website analytics, purchase history, CRM data, and social media posts, to identify patterns, spot anomalies, and generate forecasts with surprising accuracy. But it is not just about the numbers. These tools are translating raw data into clear, actionable insights, helping researchers and strategists move from descriptive data (“what happened”) to prescriptive guidance (“what to do next”). The integration of behavioral data and AI is at the heart of predictive market research, allowing for faster and more accurate decisions.
Of course, this doesn’t mean traditional methods are obsolete. Surveys still play a critical role in understanding motivations, emotions, and the “why” behind consumer actions. They’re particularly useful in early-stage product development, brand perception studies, and testing creative concepts. But increasingly, surveys are being complemented or even preceded by predictive techniques that shape where and how questions are asked.
There’s also a shift in how research teams are structured. We're seeing data scientists working alongside qualitative researchers, blending statistical modeling with human-centered design thinking. The most forward-thinking research departments aren’t picking one method over the other. Instead, they are integrating them to get a more complete, nuanced view of the market.
But with all this advancement comes a new responsibility. Predictive analytics depends on data, and a lot of it. Market researchers must now be more mindful than ever about how that data is collected, stored, and used. Data privacy laws are tightening, and consumers are becoming more aware of how their information is being tracked. Trust and transparency are quickly becoming just as important as accuracy.
At its core, market research is still about understanding people. That hasn’t changed. What has changed is the how. Instead of relying solely on consumers to tell us what they think through a form or a phone call, we now have the tools to listen to what their actions are already saying. And in many ways, those actions tell a more complete story.
We’re entering the era of predictive market research, where data doesn’t just describe what happened, it guides what to do next. For researchers, analysts, and business leaders alike, the question isn’t if they should adapt, but how fast they can.
Want to learn more about how we're using AI? Check out what we're doing with ThinkNow Synthetic?
I attended the Quirk's Los Angeles Market Research Event last week, and one thing became apparent: AI is coming for Market Research. I sat through presentations and sales pitches on AI Qual moderation, AI Co-Workers, AI Social Media Monitoring, AI-assisted Survey Creation, Data Analysis, and Report Writing. Towards the end of the conference, I wondered if next year we would all send our AI-enabled robot doppelgangers to listen to AI presenters discussing whether humans were still necessary in consumer research.
That’s not so say that I wasn’t impressed with some of the things AI is capable of doing. AI is revolutionizing market research by streamlining processes and providing faster, more efficient insights. Here are some of the major benefits:
Despite these advancements, AI has limitations that market researchers must acknowledge:
While AI promises many advantages, rapid adoption without careful oversight presents risks:
AI is undeniably transforming market research, but it should be viewed as a tool to enhance, rather than replace, human expertise. Running forward too quickly risks running into a dead-end. The best approach is a hybrid model, where AI handles time-consuming tasks while human researchers focus on interpretation, storytelling, and strategic decision-making.
As AI continues to evolve, the key to success will be striking the right balance—leveraging its strengths while mitigating its risks. Market researchers who adapt, upskill, and find ways to integrate AI effectively will be the ones leading the industry, not just observing its transformation. And hopefully, our AI doppelgangers will decide that humans are useful and nice to have around after all.
Ready to leverage AI in your market research? Find out how ThinkNow Synthetic can work for you.