Artificial Intelligence

The Patient, the Physician and the Machine

By Noah Pines

The Numbers Behind AI's Growing Role in the Patient Journey, and What They Mean for Healthcare Marketers

Recently, my research colleagues and I here at ThinkGen were commissioned to conduct a patient journey study on behalf of a biotechnology company, and I had the opportunity to personally moderate many of the preliminary pilot interviews with patients and care partners. The condition in question affects people across the age spectrum, but something particularly interesting emerged in my conversations with patients in their twenties and thirties. When describing the earliest stages of their diagnostic journeys, many spoke about consulting generative AI platforms almost as naturally as previous generations might have mentioned calling their family doctor or even consulting "Dr. Google." AI, it seemed, has become for many people the front door to healthcare, and rather more prominently than Dr Google or WebMD had ever managed.

Over the years, I have interviewed thousands of patients and caregivers across numerous therapeutic categories, and I have witnessed the evolution of how people investigate symptoms, seek explanations and make decisions about their health. For much of the internet era, the typical journey began with a search engine, followed by visits to familiar medical websites such as WebMD, Johns Hopkins, Cleveland Clinic, Mayo Clinic, etc. Within just the past year or two, an increasing number of patients appear to be dispensing with that preliminary expedition and beginning a conversation with ChatGPT, Claude, or Gemini instead. The distinction is important: rather than searching for information, they are engaging in a dialogue, asking follow-up questions, inputting personal context and receiving explanations that appear tailored to their circumstances.

My observations are, of course, qualitative rather than nationally representative. But recent research from KFF and Pew Research Center suggests that something significant is happening, while a major new initiative announced just yesterday (11/15/26) by Apple at HLTH 2026 promises to examine the phenomenon in considerably greater depth. Taken together, these developments raise an important question for those of us working in pharmaceutical marketing, market research and commercial strategy. Are we witnessing not simply the arrival of another health information channel, but the emergence of an entirely new patient journey?

‍

From Search Engine to Conversational Companion

The numbers are becoming difficult to ignore. According to a March 2026 KFF survey of 1,343 American adults, 32% had used AI chatbots for information or advice concerning their physical or mental health during the preceding year. This included 29% seeking information about physical health and 16% about mental health or emotional wellbeing. Generative AI has already reached approximately the same level of reported use as social media for health information, although healthcare professionals (HCPs) and conventional internet searches remain considerably more widely used.

The generational differences are particularly revealing. Among adults aged 18 to 29, 36% reported using AI for physical health information, while 28% had consulted it about mental health, compared with just 8% of adults aged 50 and older for the latter. Younger consumers are not necessarily abandoning traditional medical information sources, but they are incorporating conversational AI into their information-seeking behavior at an impressive rate. For an industry accustomed to studying generational differences in media consumption, treatment preferences and healthcare engagement, this represents another potentially consequential divide.

What consumers are asking is equally interesting. KFF found that 27% of all adults had used AI to investigate symptoms or obtain general information about health conditions, while 19% had sought explanations of medical tests, laboratory results or diagnoses. Another 19% had used it to understand or compare treatment options, and 16% had sought assistance deciding whether to consult a physician. These findings indicate that AI is already influencing some of the most consequential moments in the patient journey, from making sense of an unfamiliar symptom to weighing treatment options and deciding whether a visit to the doctor is warranted.

‍

Convenience, Curiosity and the Economics of Access

Why are consumers turning to AI? The most obvious explanation is also the most powerful: convenience. Among those who had used AI for health information, 65% cited the desire for quick or immediate answers as a major reason, while 41% wanted information before deciding whether to see an HCP. Another 36% valued the privacy of asking sensitive questions, and 28% had received medical test results before having an opportunity to discuss them with a clinician. AI is available around the clock, does not require an appointment and rarely appears irritated by a patient's seventeenth follow-up question.

There is, however, a less comfortable dimension to this enthusiasm. Among AI health users under 30, 38% cited the absence of a regular HCP or difficulty obtaining an appointment as a major reason for consulting AI, while 29% pointed to the cost of medical care. Among users with annual household incomes below $40,000, 32% identified affordability as a major motivation. In other words, AI adoption is not simply a triumph of technological sophistication; it may also reflect persistent shortcomings in healthcare access. A chatbot that answers immediately is understandably attractive when the alternative involves an expensive consultation several weeks hence.

Interestingly, consumers do not necessarily regard AI as a substitute for medical expertise. Only 18% of AI health users cited confidence in its reliability relative to an HCP's advice as a major reason for using it, yet 92% of those consulting AI about physical health reported being at least somewhat satisfied with its responses. Evidently, consumers need not regard AI as a medical authority to find it valuable. They appear increasingly comfortable relying on a technology whose limitations they may recognize, but whose convenience and practical utility are difficult to resist.

‍

A Companion at Every Stage of the Patient Journey

For healthcare marketers, the most consequential development may be the expanding role of AI across the entire patient journey. During symptom recognition, AI can help consumers comprehend and organize unfamiliar sensations and explore possible explanations. During diagnosis, it can translate medical terminology, interpret laboratory results and suggest questions for HCPs. During treatment evaluation, it can explain therapeutic alternatives, mechanisms of action, potential adverse effects and the practical implications of different approaches. During ongoing management, it can help patients understand monitoring requirements, prepare for follow-up appointments and navigate the everyday complications of living with chronic illness.

Care partners may find these capabilities particularly valuable. A spouse helping manage a complex medication regimen, an adult child interpreting a parent's laboratory results or a parent researching a child's newly diagnosed condition can engage with AI repeatedly as circumstances change. Unlike a static medical website, a conversational system can adapt and incorporate information provided earlier in the discussion as well as respond to increasingly specific questions. The resulting interaction may feel less like consulting a reference library and more like having a reasonably knowledgeable assistant permanently available, albeit one whose confidence can occasionally exceed its competence.

This introduces a new challenge for conventional patient journey mapping. In the past, we have traditionally identified information-seeking moments, influential channels, HCP interactions and decision points as relatively distinct junctures in the journey. AI potentially cuts across these boundaries, accompanying patients from initial uncertainty through diagnosis, treatment selection and long-term management. The important question is no longer simply where patients obtain information, but how a continuing dialogue shapes their understanding, expectations and subsequent behavior.

‍

Apple Decides to Investigate

At HLTH 2026, Apple announced an initiative that could substantially deepen our understanding of these interactions. The Apple Health AI Usage Study, conducted in collaboration with OpenEvidence, aims to enroll up to 200,000 American participants aged 21 and older. Through a dedicated application, participants will be able to ask health questions, examine responses grounded in peer-reviewed medical literature and evaluate the usefulness and clarity of the information they receive. They may also choose to incorporate selected personal health information, including activity, sleep, heart-rate measurements and laboratory results.

What makes the initiative particularly interesting is its emphasis on behavior rather than merely adoption. As described in FierceHealthcare's coverage, the study will examine trust, comprehension, responses to AI-generated information and the ways consumers subsequently interact with healthcare professionals. Participants will complete recurring surveys, potentially allowing researchers to investigate whether AI helps them prepare for medical appointments and influences what actually happens afterwards. The study has only just been launched, so its findings remain prospective, and eligibility requirements involving compatible iPhones will limit its representativeness. Nevertheless, the research questions are precisely those that pharmaceutical market researchers should be asking.

Apple's partnership with OpenEvidence is also noteworthy. By drawing upon established medical literature and clinical references, the initiative reflects growing recognition that conversational fluency is not a substitute for scientific credibility. Yet even medically grounded responses can be misunderstood, selectively interpreted or applied inappropriately to individual circumstances. Understanding the difference between the technical accuracy of an answer and a patient's ability to act safely upon it may prove one of the most important contributions of the research.

‍

The Curious Paradox of Trust

There is an intriguing contradiction in Americans' attitudes towards healthcare AI, one I've written about in previous essays on this topic. An October 2026 Pew Research Center survey of 3,488 adults found that 60% would feel uncomfortable with their own physician using AI for activities such as diagnosis and treatment, while only 30% expressed comfort. Yet substantial proportions were receptive to AI performing particular healthcare tasks under human supervision. For example, 53% were open to AI explaining laboratory results with a human conducting the final review, while 60% supported similar oversight when AI analyzed medical scans.

The apparent contradiction becomes more interesting when one considers younger adults. Although they are among the most enthusiastic users of AI for health information, 45% of Americans under 30 believed that AI use in diagnosis and treatment would worsen patient outcomes, compared with 31% expecting improvement. Across the population, 47% anticipated that AI would damage the patient-physician relationship, while 51% expected it to worsen the security of personal health information. Consumers, it appears, are perfectly capable of embracing a technology personally while remaining suspicious of its institutional deployment. The public is not always consistent, but neither is the healthcare system.

Privacy presents another contradiction. KFF found that 77% of Americans expressed concern about the privacy of medical information supplied to AI systems, yet 41% of AI health users reported having uploaded personal medical information, including test results or physician notes. Convenience and curiosity are evidently powerful competitors to caution. For healthcare companies, this tension reinforces the importance of transparency, responsible information practices and the recognition that consumer enthusiasm should not be mistaken for unconditional trust.

‍

What This Means for Pharmaceutical Marketing and I&A

The implications for pharma, biotech and medtech companies extend well beyond conventional digital marketing. Patient journey research must increasingly investigate how AI enters the decision-making process, which questions consumers ask, how they evaluate responses and whether those interactions influence subsequent clinical conversations. Researchers need to deeply explore actual AI-assisted experiences rather than relying exclusively on traditional questions about websites, search engines and patient communities. Quantitative studies should distinguish between AI use for general education, diagnosis-related questions, treatment comparisons and ongoing disease management, since these activities carry different implications for patient behavior.

Medical information and content strategies also deserve reconsideration. In a world where consumers increasingly receive synthesized answers rather than lists of websites, the discoverability, accuracy and clarity of authoritative medical evidence become strategically important. Manufacturers should consider how high-quality, balanced and appropriately contextualized information about diseases and treatments can be understood and represented by AI systems, while respecting regulatory boundaries and the independence of clinical decision-making. This is not simply search-engine optimization with a more fashionable acronym; it is a question of how scientific information reaches consumers through an increasingly influential intermediary.

Perhaps most importantly, our industry should resist the temptation to treat AI as merely another channel to be added to an already crowded communications plan. Its potential significance lies in its ability to participate repeatedly in the consumer's reasoning process, influencing how symptoms are interpreted, questions are formulated, options are compared and clinical encounters are approached. We will increasingly need research methodologies capable of understanding these interactions over time, including where AI improves patient understanding and where it introduces confusion, misplaced confidence or misinformation.

The patient journey has never been as linear as our PowerPoint diagrams depict. But conversational AI may make those familiar diagrams look particularly antiquated, introducing an assistant that can accompany patients and caregivers across stages we have traditionally examined separately. KFF has demonstrated that adoption is already substantial, Pew has illuminated the complicated boundaries of public trust, and Apple's ambitious new study may help reveal what consumers actually do with the information they receive. For those of us whose business depends upon understanding patients, the task is not simply to monitor the growth of AI, but to reconsider the journey itself. Dr Google has not retired, but he may increasingly find himself relegated to the waiting room.