From Interview #87
With Rajiv Haravu
Rajiv Haravu outlines the complex challenges in healthcare data normalization, emphasizing variability in clinical documentation and the real-world consequences of non-standardized data. He explains how differences in terminology, context, and documentation style can lead to information loss, affecting patient care, research, and public health. Haravu also discusses 'definition decay'—how medical terminology and meaning evolve over time—and highlights IMO Health’s approach of constant surveillance, expert curation, and regular content updates to maintain accuracy.
From Interview #87
With Rajiv Haravu
Rajiv Haravu, SVP of Product Management at IMO Health, explains why medical code sets remain essential despite the rise of large language models. While generic LLMs can generate answers, they often lack the specialized context needed for accurate clinical coding and are prone to hallucinations. By combining LLMs with proprietary medical content, structured code sets, and editorial policies, IMO Health significantly improves AI accuracy. Haravu also explores the potential of knowledge graphs to link diagnoses, labs, medications, and procedures, as well as using AI agents to streamline dictionary migrations in EHRs.
From Interview #83
With Pelu Tran
Pelu Tran, CEO of Ferrum Health, outlines why AI adoption in hospitals remains slow despite the technology’s readiness. The barriers lie in integrating modern, cloud-based AI into legacy systems, navigating multimillion-dollar onboarding processes, and addressing strict patient data governance. Tran warns that most AI tools underperform in real-world conditions and require continuous monitoring for bias, drift, and workflow impact. He advises hospitals to view AI as a lifecycle rather than a point solution, building governance frameworks to manage performance, safety, and cost over time.
From Interview #85
With Jason Alan Snyder
Jason Alan Snyder explores two possible futures for big data in healthcare: a dystopia where patients lose all agency and a utopia where digital twins empower individuals. In the dystopian scenario, consent is reduced to a checkbox, algorithms trained on flawed data determine access to care, and patients have no visibility into decisions. In contrast, the utopian vision puts patients in control of their data, with digital twins serving as personal advocates that guide care, warn early, and even generate income. Snyder argues that achieving this future requires courage, will, and—most importantly—better healthcare data quality as the foundation for trustworthy AI tools.
From Interview #76
With Dr. Debra Patt
Dr. Debra Patt, Chair of the AI Taskforce at the American Society of Clinical Oncology (ASCO), discusses the organization’s guidelines for responsible AI use in cancer care. While ASCO recognizes AI’s potential to improve decision-making, Patt emphasizes transparency, bias awareness, and keeping physicians central in care delivery. She illustrates with a clinical example how AI recommendations may overlook patient-specific factors, reinforcing the need for human oversight. AI can offer valuable nudges or decision-support suggestions, but ultimate care choices must remain personalized and patient-centered.
From Interview #88
With Dr. David Fajgenbaum
Dr. David Fajgenbaum explains how Every Cure is democratizing drug discovery by releasing an open-source database of AI-generated drug-disease scores. Using advanced knowledge graphs, the platform evaluates all 4,000 approved drugs against every known disease, creating over 75 million rankings. While Every Cure uses these scores internally to prioritize research and clinical trials, they will soon make the entire dataset publicly available. This resource will allow scientists, clinicians, and even patients to identify promising drug candidates for conditions ranging from Parkinson’s to rare diseases, accelerating global collaboration in medical innovation.
From Interview #78
With Bob Battista
Bob Battista draws parallels between liability in self-driving cars and AI in healthcare, suggesting that true transformation will come when patients can self-drive their own care. He argues that by giving patients access to their health data and enabling AI to process it, individuals can make informed decisions about treatment options without overburdening clinicians. This shift could reduce liability concerns while also accelerating the sharing of patient knowledge, allowing newly diagnosed individuals to start their care journey armed with the best available insights.
From Interview #78
With Bob Battista
Bob Battista explains how AI can transform healthcare decision making by creating personalized treatment guidelines in real time. By integrating patient chart data with the vast body of published clinical evidence, AI can identify the most effective treatment paths for each individual. While access to proprietary data inside pharmaceutical companies could further enhance these capabilities, even existing public data offers tremendous potential. Battista emphasizes that applying best practices earlier in the care process could improve outcomes and save significant costs, especially if tailored to at-risk patients.
From Interview #80
With Dr. Azra Raza
Dr. Azra Raza discusses the transformative potential of continuous AI-driven health monitoring to detect diseases like cancer before symptoms arise. She describes implantable devices capable of screening the entire bloodstream every 18 days for abnormal cells and molecular biomarkers, enabling intervention at the earliest possible stage. While FDA approval for these devices is pending, point-of-care technologies already allow high-frequency testing for illness signatures at home. Raza highlights the urgency of funding these innovations, emphasizing that early detection could shift healthcare from reactive treatment to true prevention.
From Interview #77
With Tom Neyarapally
Tom Neyarapally, CEO of Archetype Therapeutics, explains why drug discovery remains slow despite major technological advances. He highlights that while access to molecular profiling, patient clinical data, advanced AI/ML tools, and scalable compute power have transformed the field, traditional workflows remain fragmented. Neyarapally advocates for a patient-centric approach—starting with well-characterized patient cohorts and using AI to screen drugs that can move them from illness toward wellness. This method, coupled with repurposing existing safe drugs, can dramatically shorten proof-of-concept timelines.
From Interview #84
With Dr. Alister Martin
Dr. Alister Martin, CEO of A Healthier Democracy and Assistant Professor at Harvard Medical School, discusses how AI can improve emergency room efficiency by identifying high-need patients—often called 'frequent flyers'—and connecting them with essential social services. Through a process called benefit stacking, AI streamlines applications for multiple assistance programs, addressing root causes like housing instability and utility costs. Martin cites evidence from randomized controlled trials showing that small, targeted financial support can dramatically reduce ER visits, cut Medicaid costs, and relieve hospital financial strain.
From Interview #83
With Pelu Tran
Pelu Tran, CEO of Ferrum Health, addresses the pressing issue of patient data security in the era of AI. He explains why hospitals overwhelmingly prefer AI systems to run within their own controlled environments—either on-premises or in their own cloud—rather than in vendor-controlled clouds. Pelu outlines the risks of vendor environments, including breaches, unauthorized model retraining, and misuse of aggregated data. He also discusses the regulatory barriers, such as FDA requirements that limit adaptive model updates, and how new open-source models could reshape secure AI deployment in healthcare.
From Interview #79
With Dr. Nigam Shah
Dr. Nigam Shah, Co-founder of Atropos Health and Chief Data Scientist at Stanford Healthcare, examines the sustainability challenges in AI development for healthcare. He explains that while AI in medicine has existed for decades, the current academic-centric development practices are not suited for scaling from research to real-world applications. The conversation highlights the cost, time, and regulatory complexities, as well as the need for localized and continuous model validation to maintain performance.
From Interview #79
With Dr. Nigam Shah
In this conversation, Dr. Nigam Shah, Co-founder of Atropos Health and Chief Data Scientist at Stanford Healthcare, challenges how the healthcare industry applies AI. He warns against the 'Turing trap'—using AI to replicate tasks clinicians already perform—instead of leveraging it to reimagine workflows. Shah shares examples of AI applications that could meaningfully expand patient access, like triage tools that reduce unnecessary visits. He also highlights how misaligned incentives block data sharing, which limits innovation despite having the necessary technology. This discussion addresses both healthcare tech trends and the urgent need for better AI integration strategies.
From Interview #79
With Dr. Nigam Shah
Dr. Nigam Shah, Co-founder of Atropos Health and Chief Data Scientist at Stanford Healthcare, shares his framework for assessing AI’s role in improving patient care. He describes the 'firm assessment' process, which evaluates expected benefits, feasibility, and incentives before deploying models. Shah also critiques current AI-in-healthcare research for lacking real-world EHR data and focusing solely on accuracy metrics, advocating instead for task-specific evaluation and uncertainty measures. He emphasizes that impactful AI requires starting with the right clinical question, selecting datasets fit for that question, and integrating evaluations into the organization’s responsible AI lifecycle.
From Interview #81
With Krish Ramadurai
Krish Ramadurai, Partner at AIX Ventures, outlines his strategy for selecting drug assets in a crowded and competitive pharma landscape. He emphasizes starting with deep conversations with key opinion leaders (KOLs) to identify areas that have broad consensus among industry experts. Ramadurai recommends diversifying assets to avoid over-reliance on a single program, advocating for at least one clinical-stage asset supported by multiple preclinical backups. He also stresses the importance of founder engagement and having clear out-licensing strategies for assets that don’t align with the founder’s passion or company focus.
From Interview #81
With Krish Ramadurai
Krish Ramadurai, Partner at AIX Ventures, shares insights into the most promising areas of AI-native health tech. From scalable AI diagnostics that combine multiple tests into one platform to full-stack biotech models that bypass inefficient CRO workflows, Ramadurai outlines how automation is reshaping pharma. He highlights opportunities in precision oncology, immunology, and lab automation, as well as the challenge of integrating disparate lab systems. The conversation points to a potential reinvention of big pharma, driven by efficiency gains, in-house automation, and strategic AI deployment.
From Interview #88
With Dr. David Fajgenbaum
Dr. David Fajgenbaum, Co-founder of Every Cure, explains how AI is accelerating the search for treatments for rare diseases through drug repurposing. By scanning all existing drugs across all diseases, AI dramatically increases the chances of finding matches for conditions that otherwise have no effective therapies. While any individual rare disease may have low odds of an existing treatment, the cumulative approach across all conditions reveals many untapped opportunities. Fajgenbaum emphasizes the moral obligation to connect patients with available therapies—especially when they already exist in local pharmacies—rather than forcing them to wait years for new drug development.
From Interview #80
With Dr. Azra Raza
Dr. Azra Raza, Professor of Medicine at Columbia University, argues that stage one cancer detection is already too late. She critiques current screening methods—which generate millions of false positives and cost billions annually—and advocates for continuous health monitoring from birth. Raza envisions implantable stents equipped with sensors and chips that can detect abnormal cells in real time, transmitting alerts to a patient’s phone. Coupled with AI and machine learning, this approach could enable detection at the 'first cell stage,' revolutionizing cancer prevention and patient outcomes.
From Interview #84
With Dr. Alister Martin
Dr. Alister Martin shares how AI is transforming emergency medicine by tackling two of the ER’s biggest challenges: documentation and rapid patient assessment. He describes how ambient AI systems automatically capture doctor-patient conversations and generate accurate clinical notes, freeing clinicians from hours of charting. In high-pressure moments, large language models can instantly summarize a patient’s medical history, allowing physicians to make informed decisions within seconds. These tools not only save time and reduce burnout but also improve care quality by ensuring no critical detail is overlooked.
From Interview #77
With Tom Neyarapally
Tom Neyarapally explains how Archetype Therapeutics uses AI to revolutionize cancer research by drastically reducing drug discovery timelines. By merging established genomic classifiers—tests that predict cancer outcomes—with cutting-edge generative AI, the company can virtually screen billions of molecules in a single day. This process targets multiple genes simultaneously, identifying compounds most likely to improve patient survival. By eliminating the need to physically test low-probability candidates, this approach saves time, reduces costs, and accelerates the development of life-saving therapies.
From Interview #84
With Dr. Alister Martin
Dr. Alister Martin, CEO of A Healthier Democracy and an emergency physician at Harvard, explains what AI is used for in healthcare today: solving real, upstream pain points. Through Link Health, his team uses large language models to connect Medicaid patients to more than billion in unspent federal and state aid, reducing avoidable ER visits and addressing social emergencies like food and housing insecurity. The conversation explores how AI in nonprofits can streamline complex benefit navigation and support systems-level improvements aligned with ai in healthcare policy. For health leaders, this is a practical path to lower costs and improve outcomes by meeting patient needs before they become medical crises.
From Interview #78
With Bob Battista
Healthcare leaders ask why proven medicines still aren’t widely reused. In this short conversation, Bob Battista explains that the core barrier to drug repurposing isn’t technology—it’s policy and incentives. While AI drug repurposing and real‑world data can surface new indications, the most valuable knowledge remains locked inside pharmaceutical organizations and constrained by regulatory risk and reimbursement dynamics. Battista outlines how safe‑harbor data sharing and new financial instruments could let companies support niche indications without eroding primary markets, accelerating access for patients and clinicians. He also highlights the untapped insights from physicians’ off‑label use and patient experience—critical signals the healthcare system rarely aggregates. If you work in market access, clinical operations, or digital health, this is a clear roadmap to move the drug repurposing market from potential to practice.
From Interview #76
With Dr. Debra Patt
Dr. Debra Patt outlines the many ways AI can address the financial burden on healthcare systems, especially in oncology. She highlights AI-powered patient education tools that prevent costly ER visits, AI-enabled drug discovery methods like AlphaFold that streamline development, and smarter clinical trial designs that reduce waste. By targeting precise molecular mechanisms, AI offers the potential for more effective, less toxic treatments. Patt emphasizes the dual benefit of improving patient outcomes while being responsible stewards of healthcare spending.
From Interview #76
With Dr. Debra Patt
In this insightful discussion, Dr. Debra Patt explores the nuanced balance between patient privacy, data monetization, and the transformative role of AI in healthcare. She highlights that while individual patient records hold limited value, aggregated, de-identified data can drive significant medical advancements, such as expanding drug indications through real-world evidence. Dr. Patt also addresses the challenges posed by electronic health record (EHR) systems, noting their limitations as billing-focused tools that often fail to capture accurate clinical data in real time. For AI to truly revolutionize healthcare data use, she argues, both technology and clinical workflows must evolve together.
From Interview #86
With Dr. Euan Ashley
Dr. Euan Ashley, Chair of the Department of Medicine at Stanford, challenges the outdated criteria for medical school admissions. He argues that while chemistry and physics have long been gatekeepers, they fail to assess skills critical for future doctors, such as AI proficiency, adaptability, and empathy. In a world where language models, multimodal diagnostics, and real-world data will be essential in care delivery, doctors unskilled in AI risk falling behind. Ashley calls for admissions reform to evaluate candidates on the abilities that will matter most in the AI-powered healthcare environment.
From Interview #86
With Dr. Euan Ashley
In this conversation, Dr. Euan Ashley explains how wearable devices can identify atrial fibrillation (AFib), a major risk factor for stroke, before symptoms occur. By monitoring the intervals between heartbeats, these devices detect the irregular rhythm characteristic of AFib, alerting users to seek medical evaluation. Unlike traditional detection methods—which require short-term clinical monitoring—wearables provide continuous, long-term observation, making it possible to catch AFib early, even years before it might be detected in a clinic. Ashley notes that while sophisticated AI can enhance diagnostics, detecting AFib requires only identifying regular versus irregular rhythms.
From Interview #85
With Jason Alan Snyder
Jason Alan Snyder breaks down the challenges of poor data quality in healthcare, particularly the issue of data decay. He explains that while AI models depend on accurate and relevant data, much of today's healthcare information is outdated, incomplete, or biased, leading to flawed recommendations. Snyder identifies four main types of data decay—temporal, structural, contextual, and semantic—and highlights the risks of semantic manipulation in shaping medical terminology for market advantage. He advocates for clean, consent-based, and personally controlled data as the foundation for ethical, precise, and effective healthcare AI.
From Interview #85
With Jason Alan Snyder
Jason Alan Snyder tackles the complex question of health data ownership, arguing that individuals—not corporations—should control their medical information. Despite legal rights under HIPAA, patients face significant barriers to accessing their own data, which is often siloed, unstructured, and inaccessible in a useful form. Snyder notes that providers, labs, pharma, and tech platforms profit from patient data while offering little in return to its source—the patient. He calls for systems that unify fragmented data, create portability and usability, and enable individuals to profit from their health information ethically and securely.
From Interview #81
With Krish Ramadurai
Krish Ramadurai explains how AI is transforming biotech from a high-risk guessing game into a predictive science. He discusses the need for diverse, human-based datasets to avoid the pitfalls of biased models that lead to drug failure. Ramadurai highlights the value of rescuing shelved drugs using chemo-informatics to tweak their properties for renewed clinical viability, as well as investing in companies with strong data moats. By combining robust biology with advanced AI, he envisions reversing biotech’s current 95% failure rate and accelerating time to market for safe, effective therapies.
From Interview #81
With Krish Ramadurai
Krish Ramadurai breaks down the reality of AI adoption in healthcare, separating transformative applications from overhyped promises. He emphasizes that success comes from domain expertise, frictionless integration, and solving real workflow problems—not from replacing physicians. Generative AI has proven effective in clinical workflow automation, like dictation and EHR integration, but faces limitations in complex decision-making domains. Ramadurai warns against 'pilot purgatory' and stresses the importance of demonstrating clear ROI, enterprise readiness, and alignment with the needs of patients, payers, and providers.
From Interview #85
With Jason Alan Snyder
Digital twins—AI-powered models based on individual health data—are poised to transform how healthcare predicts, treats, and personalizes care. But Jason Alan Snyder, a technologist and futurist, raises powerful ethical and technical concerns about how these systems are being deployed today. Built from EHRs, lab results, wearables, and even genomics, these digital representations can simulate decisions and even replace human judgment. Yet they are often created without consent or clarity. In this clip, Snyder urges healthcare leaders to reclaim agency over the design and governance of these digital surrogates, emphasizing that the future of AI in healthcare must be rooted in truth, transparency, and trust.
From Interview #77
With Tom Neyarapally
What’s next for AI in drug discovery and development? Tom Neyarapally, CEO of Archetype Therapeutics, shares his outlook on how converging technologies—from LLMs to spatial omics—are reshaping how we discover and deliver drugs. He emphasizes that while AI tools are evolving rapidly, the integration of diverse data modalities and collaborative innovation between nimble startups and large pharma is essential. Neyarapally also touches on reducing late-stage trial failures and costs—key challenges that AI is finally beginning to address. For stakeholders focused on AI in drug development, this conversation offers both strategic insight and practical optimism.
From Interview #82
With David Norris
David Norris, a lifelong technologist and healthcare innovator, unpacks why AI—especially large language models (LLMs)—is gaining rapid traction in healthcare. With decades of progress converging in recent breakthroughs, Norris outlines how AI can now handle tasks from reading faxes to calling patients about lab results. He emphasizes the practical LLM use cases in healthcare that free clinicians from administrative burdens, allowing more time for patient care. This transformation isn't about replacing jobs—it's about restoring the human connection in medicine by shifting repetitive tasks to AI. The conversation brings clarity to the real benefits of AI in healthcare and where it's headed next.
From Interview #83
With Pelu Tran
In this insightful discussion, Pelu Tran explores the practical role of artificial intelligence in enhancing clinician productivity and navigating the widespread inefficiencies in healthcare. Drawing from Ferrum Health’s platform, Tran explains how AI enables faster diagnosis, supports care coordination, and amplifies provider capacity—particularly in overburdened fields like radiology. He also highlights the real cost of AI adoption, pointing to the significant resources required to integrate AI into existing healthcare systems. This conversation sheds light on the evolving impact of AI for healthcare productivity and reframes the debate: it’s not just about quality vs. efficiency—it’s about enabling clinicians to sustain care at scale.
From Interview #88
With Dr. David Fajgenbaum
In this compelling conversation, Dr. David Fajgenbaum describes how his team uses a platform called Matrix, powered by a biomedical knowledge graph, to scan all known relationships between drugs, diseases, and genes. The goal? To apply AI for healthcare data analysis in identifying potential treatments, some of which had been previously overlooked or discarded due to lack of profitability. One powerful example includes the identification of a TNF inhibitor for Castleman disease that led to a lifesaving intervention. Dr. Fajgenbaum emphasizes how this technology doesn't just discover new solutions but also revalidates forgotten ones, shedding light on opportunities to repurpose existing drugs. This showcases the transformative potential of AI predictive analytics in healthcare.
From Interview #80
With Dr. Azra Raza
In this compelling interview, Dr. Azra Raza challenges the traditional focus of oncology, arguing that early cancer detection should be the central pillar of cancer research and treatment. Sharing her personal journey as an immigrant and oncologist, Dr. Raza critiques the widespread reliance on animal models and late-stage therapies, advocating instead for the study of human tissue and the identification of disease at its earliest stages. Her lifelong work in hematologic malignancies, particularly myelodysplastic syndromes, underscores the urgency of this shift. By focusing on “the first cell” rather than the last, she highlights how early intervention could dramatically reduce suffering and improve survival—raising critical questions about current priorities in cancer care and the persistent challenges in cancer treatment.
From Interview #84
With Dr. Alister Martin
In this clip, Dr. Alister Martin outlines how both AI and healthcare policy can reduce the cost of care. While his organization, A Healthier Democracy, remains people-first in its approach, Dr. Martin strongly advocates for AI upskilling as essential in the modern workforce. He warns that it's not AI that will replace workers, but workers who use AI that will replace those who don't. On the policy side, Dr. Martin makes a compelling case for maintaining reimbursement pathways through Medicare and Medicaid to sustain initiatives that demonstrably lower emergency room visits and hospitalizations—highlighting the cost-effectiveness of AI in healthcare. His remarks provide actionable direction for organizations aiming to use AI for healthcare cost saving strategies.
From Interview #76
With Dr. Debra Patt
In this insightful interview, Dr. Debra Patt shares how AI applications in healthcare are transforming the daily realities of clinical care. As a breast medical oncologist and health policy leader, she describes the power of ambient AI scribe technology, which reduces administrative burdens and improves patient communication by instantly generating visit notes. She outlines specific benefits—from more accurate ICD-10 coding to enhanced decision-making in cancer care—that not only help providers get patients the right treatment faster, but also streamline billing and prior authorization. Her perspective offers a grounded, real-world view of "what are the applications of AI in healthcare" today, especially for oncology practices navigating high patient loads and value-based care demands.
From Interview #82
With David Norris
Can AI replace doctors—or is it here to restore what’s been lost in medicine? David Norris, CEO of Affineon Health, tackles this provocative question with depth and clarity. Far from eliminating physicians, Norris argues that AI solutions are essential for addressing physician burnout, streamlining administrative tasks, and reviving the human relationship between doctors and patients. He illustrates how AI, like Affineon’s $2.50/day AI assistant, can triage clinical inboxes, flag clinically significant lab trends, and bring 10 years of patient history into a three-second review—freeing doctors to focus on what matters most. This conversation is a must-watch for anyone curious about how AI is reshaping the future of care.