Aashima Gupta

Those “direct-to-physician” CDS AI tools like OpenEvidence and DoximityGPT? They're now going after health systems too. My 5 thoughts on how this will all play out: First, the gist of what was announced: → Sutter Health will integrate OpenEvidence with Epic, allowing physicians to access the AI medical search engine directly within the patient chart → UC San Diego Health is rolling DoximityGPT - their AI medical search engine - across the organization, while also making its AI scribe tool available My 5 thoughts on all this: 1/ This move to compete in the enterprise was 100% inevitable. Especially for OpenEvidence that just raised $250M at a $12B valuation, the only way to justify that valuation is to keep rapidly increasing revenue growth - and they can’t do that without increasing their Total Addressable Market beyond individual physicians. If they are truly already used by 40% of US physicians, then revenue growth is going to plateau if they don’t grow into adjacent markets. 2/ Is the Sutter Health integration of OpenEvidence and Epic just an embedded view for easy access? Or is there true data integration where a doctor can ask a question and OpenEvidence already has full context about the patient, and thus can produce a personalized answer? From the press release it sounds like just an embedded view for now, but full data integration for personalized CDS - the holy grail - seems inevitable. We know Evidently, AvoMD and others have already done this so it’s a matter of when, not if. 3/ What will be the business model? I suspect it will switch from an ad-model to a SaaS model. Health systems typically avoid ads to doctors and will also require more significant professional services for integrations, training, account management, etc and thus also cost more. 4/ How will this all work when EHRs are pushing doctors to use their own AI clinical assistants (e.g. Epic Art)? Unlike the growing competition between EHRs and AI scribes, I suspect Epic will allow companies like OpenEvidence, Doximity and UpToDate to just directly connect into Art - so doctors can access the same CDS experience. Chat itself is now a universal form factor, and as long as you’re selling licenses, the CDS companies don’t care what the access point is. As for Epic, they want to be a “software factory” and not a clinical guidelines/content factory - so I predict Epic will release a Toolbox category for CDS AI to integrate with Art. 5/ This will put major pricing pressure on UpToDate - however, UpToDate will have a head start on integrating with EHR AI assistants like Epic Art, given their longstanding partnerships with the EHRs. If I were UpToDate, I would be doing a mad dash to be the first to get AI CDS into Art and figure out a revenue share that incentivizes Epic to promote their product.

151

50 Comments

My 7 biggest takeaways from The Permanente Medical Group, Inc. experience using AI Scribes after 1 year. This is particularly interesting given the switch from Nabla to Abridge after the decision to go enterprise-wide with AI Scribes by Kaiser Permanente. 1️⃣  72% of physicians used the AI Scribe at least once among a sample of 10,000+ physicians who were given access Also, at the end of 2024 the adoption curve was still trending up, so it may not have plateaued just yet. 2️⃣ 34% of physicians used AI Scribes in at least 100 patient encounters… but that means the other 38% of physicians who tried it only used it for an average of less than 0.38 patient encounters per work day (< 100 encounters / 260 work days) In other words, usage is driven by the top cohort of users. This is what will drive the argument from health systems from why the pricing model should not be “per user” and why the enterprise license should be at a fraction of the original “per user” costs. Ambient vendors will argue that the “per user” pricing already baked in this distribution of usage among possible users, and this is still a fraction of what hiring human scribes for every clinician would cost. 3️⃣ Highest physician adoption was among mental health (42%), primary care (32%) and emergency medicine (32%), based on % adoption by clinicians in each specialty. No real surprises here. 4️⃣ The top 33% of physicians who tried the AI Scribe (“High Users”) were responsible for an overwhelming 89% of the total usage of the AI scribe. In fact, High Users also were the ones who needed it the most - as pre-AI Scribe they tended to be physicians with more time spent in the EHR per appointment and fewer patient visits per day. Not provided: did High Users end up now doing more patient visits per day than before? 5️⃣By the end of 2024, High Users spent even less time in notes than Low Users and Non-Users (i.e. didn’t try AI scribes even once) → High Users went from 4.5 minutes/note to 2.5 minutes/note → Low Users went from 3.75 minutes/note to 2.9 minutes/note → Non-Users went from 3.75 minutes/note to 3.4 minutes/note Will data showing that maybe they could save 30 seconds per note motivate the Non-Users to become Users? I’d guess No. 6️⃣ 57% of patients experienced no difference in time physicians spent looking at a computer screen (caveat: small n=118 sample) This was surprising. Begs the question: were some physicians weren’t looking at a screen much normally during a visit, or whether they just couldn’t shake the habit? 7️⃣ Adoption of AI Scribes were mostly agnostic to the vendor Although the total physician adoption essentially doubled after switching vendors, the team felt that this was related more to: → The positive experience of Vendor 1’s pilot → Increased awareness and support provided over 2024 → Implementation of EHR integration with Vendor 2 Hard to tell how KP truly felt about this or if they were just being nice. Full article in the comments 👇

250

53 Comments

I’m proud to share that Kaiser Permanente has launched a new public policy page focused on artificial intelligence. This page outlines our commitment to the responsible use of AI—anchored in equity, safety, and transparency—and features two new videos that highlight how we’re using AI to improve care and operations across our organization. It also offers clear, actionable policy recommendations for lawmakers who share our goal of building public trust in AI.   We believe that thoughtful policy can help ensure AI serves the public good. I invite you to explore how we’re approaching this work—and how we can shape the future of health care together. https://lnkd.in/giZPf2Yd   #ResponsibleAI #HealthPolicy #KaiserPermanente

472

37 Comments

There's a new question in every health tech buying cycle that didn't exist 18 months ago: "Is this already on the roadmap of our foundation model vendor?" As we discussed in our 2026 healthcare predictions, we are entering a moment where the end user, clinicians and operators included, can increasingly build software themselves using agentic coding environments. In practice, this shows up when a CIO asks whether a proposed point solution is meaningfully different from what their existing foundation model could deliver with internal prompts, guardrails, and a few weeks of engineering time. This shift is quietly reshaping the competitive calculus. Startups selling into healthcare enterprises no longer compete only with the EHR vendor’s roadmap. They also compete with what a foundation model provider can plausibly ship for that customer, often faster and at lower marginal cost. Importantly, while the moat around building software is eroding, the moat around running software in healthcare is not. Healthcare enterprises are not set up to independently handle deep workflow integration, continuous iteration, reliability at scale, and governance across clinical, legal, and regulatory dimensions. That operational gap is where durable companies are still built. We expect this dynamic to introduce real friction into buying cycles for point solutions with thin moats. At the same time, it expands opportunity for companies with defensible data assets, deep workflow entrenchment, and true platform potential. Artera and Atropos Health are two strong examples from Breyer Capital's portfolio, both highlighted in the article. Special thanks to Brian Gormley for featuring our perspective in today's The Wall Street Journal piece on OpenAI and Anthropic's entry into healthcare. Jim Breyer Daniel Breyer Andre Esteva, PhD Brigham Hyde

272

29 Comments

Three incredible days at #JPM2026. Same conversation with 10+ health system execs: "We're building our AI strategy. Evaluating vendors. Need to pick the right partners." This sounds familiar. 2009-2012: The EHR decision. After the HITECH Act, health systems spent 18-24 months evaluating Epic vs. Cerner. Consultants. Committees. RFPs. Multi-year implementations. The logic made sense: EHRs were monolithic platforms with huge switching costs. You had to get it right. 2025: The AI decision. I'm watching health systems do the same thing. "AI steering committees." "Strategic AI partner" RFPs. Trying to pick winners. But there's a critical difference: EHRs were a single, massive platform decision. AI in healthcare is dozens of different problems that need dozens of different solutions. Here's what health systems are missing: They're looking for "the AI vendor" when they should be building "an AI portfolio." There's no single AI platform that solves:   → Documentation burden (ambient notes) → Radiology throughput (workflow optimization) → Clinical trials matching → Revenue cycle optimization (across specialties) → Clinical decision support These are different problems. Different technologies. Different vendors. Different timelines. Some need FDA clearance and 24-month validations. Some can pilot in 90 days. Some integrate with Epic. Some replace parts of Epic. You can't solve all of these with one "strategic AI partnership”....no matter that Anthropic is saying right now. <ducks> The AI landscape is evolving too fast for monogamous vendor relationships. 6 months ago: GPT-4 for clinical documentation. Today: Ambient listening, agentic workflows, multimodal imaging analysis. 6 months from now: Something else entirely. That's why you need exposure across the whole landscape—not a bet on one vendor. The mistake I see health systems making: Approach 1: "Strategic Partner" Strategy Pick one AI vendor → Enterprise deal → 3-year commitment → Hope they solve everything Result: You're locked in. Technology moves on. You're stuck in vendor negotiations while competitors move faster. Approach 2: "Portfolio Strategy" Deploy capital across multiple companies → Take equity → Co-develop solutions → Maintain optionality Result: You capture upside. You shape products. You're first to new capabilities. You have relationships across the ecosystem. After 20 years building innovation platforms at JPMorgan, Mousse Partners, Blueprint Health, Junto Health, Atlantic Health: In fast-moving markets, portfolio strategies beat single-vendor bets. Healthcare AI is moving fast. The question isn't "Which AI vendor should we pick?" The question is: "How do we build an AI portfolio that gives us the most optionality" If you're thinking about this, let's talk. 📩 douglas.m.hayes@gmail.com ☎️ 215-620-2032 #healthcareAI #digitalhealth #venturecapital #innovation

293

30 Comments

Doximity just released a free AI scribe for US clinicians, accelerating commoditization/the race to the bottom. Their product won’t integrate into the EHR like Abridge, but it’ll be interesting to see how much it impacts their ability to acquire customers, drive usage and maintain retention. This is the start of a trend that will continue across every vertical market. Counterpositioning with free products to undercut upstarts and incumbents alike. Buckle up. Your margin is someone else’s opportunity.

53

5 Comments

Microsoft just posted a Chief Nursing Product Lead role. About time. For years, healthcare tech was built for physicians. Everything optimized around the doctor's workflow. Meanwhile nurses inherited whatever came downstream. A new tool cuts physician documentation time by 20%? Cool. Except now nurses are manually fixing the gaps, chasing missing orders, and doing double documentation because nobody thought about their workflow. Nurses have been cleaning up after "doctor-first" products for years. But something's changing.... Over the past year, I've noticed more companies actually focusing on nursing workflows. Not as an add-on, but as the main thing. They're realizing nurses hold the system together. Microsoft hiring a Chief Nursing Product Lead is another GREAT signal. Why? Because the bottleneck in healthcare AI isn't technology it's deeply understanding workflows (including all the workarounds). Because you need someone who knows where problems are, what actually works, and who can drive adoption on the floor. Nurses are PERFECT for this. Good on Microsoft 👏 PS: Thank you Rebecca Love RN, MSN, FIEL for sharing!

642

58 Comments

Huge news from CMS for anyone building in tech-enabled care delivery & focused on chronic conditions. Yesterday, CMMI released a new outcomes-based payment model that some industry insiders are calling the biggest tailwind for digital health since COVID. It has major implications for companies managing chronic conditions, remote monitoring, virtual care, and value-based models. I spent time digging into the details with Jared Augenstein and Randi Seigel, who have been in the weeds on the ACCESS model since it dropped. Their insights helped me unpack: • How outcomes will actually be measured • Which digital health companies are best positioned • Who won’t qualify (and why) • What still isn’t clear about payment rates and targets • How this could impact PCPs, beneficiaries, and infrastructure vendors • Where the fraud and compliance risks live If you're working in digital health, chronic care, RPM, MSK, metabolic health, or infrastructure…this one is worth a read. Full breakdown here: https://lnkd.in/erSbFkd3

136

3 Comments

Great conversation yesterday at JPM on AI at the Frontier of Human Health. I had the privilege of moderating a thoughtful and practical panel with leaders who are shaping how AI is applied across the life sciences ecosystem: - Brad Lightcap, COO, OpenAI - Karen Akinsanya, President, R&D & Chief Strategy Officer, Schrödinger - Fred Manby, Co-Founder & CTO, Iambic Therapeutics - Michael Nally, CEO, Generate:Biomedicines & Partner, Flagship Pioneering - Gianluca Pettiti, EVP, Thermo Fisher Scientific We covered the full arc of AI in human health—from AI maturity and real-world impact to discovery and preclinical decision-making, clinical development at scale, and the organizational changes required to embed AI responsibly. A few themes stood out: - AI has crossed a technical inflection point, but credible impact depends on where and how it’s applied. - Better early, preclinical decisions are critical to optimizing downstream clinical development. - Scaling AI requires more than models—it demands new decision architectures, trust, and accountability. - The discussion also reinforced the importance of strong ecosystem partnerships. Our work at Thermo Fisher Scientific (& PPD, Patheon) with OpenAI and NVIDIA is helping accelerate the development of secure, scalable AI capabilities that can be responsibly applied in regulated environments. At Thermo Fisher Scientific, we’re not just adopting AI—we’re transforming how we work and how we support our customers. By combining AI automation and AI agents with deep domain expertise, we’re reimagining workflows across research, manufacturing, and clinical operations. The goal is clear: help customers bring the next innovation to patients faster, while upholding digital trust and responsible AI principles. Thank you to the panelists and everyone who joined the conversation. The future of AI in human health will be defined not by experimentation alone, but by focus, integration, and leadership commitment. #JPM #JPM2026

221

5 Comments

This week feels like a real inflection point for AI in healthcare—not because of a single breakthrough model, but because the ecosystem is finally ready. After more than a decade working at the intersection of health and tech, I can’t recall another moment when capability, incentives, and trust felt this aligned. We’ve long known healthcare is massive and structurally broken, and that health information is one of technology’s most valuable use cases. The constraint was never opportunity—it was readiness. For the first time, patients and health systems seem prepared to absorb AI in ways that are genuinely useful, responsible, and scalable. With launches spanning both consumer and enterprise, OpenAI isn’t just shipping products; it’s signaling a long-anticipated “slowly, then all at once” shift. I entered healthtech through Google more than a decade ago, working on healthcare knowledge cards in Search, early symptom-to-condition frameworks, and Google’s first HIPAA-compliant product. Even then, the signal was clear: when people have health questions, they turn to tech first—often before they ever touch a health system. Health queries made up roughly 20% of search traffic, yet users were frustrated, forced into fragmented, sideways searches to assemble answers. We even embedded telemedicine directly into Search. Despite strong teams and deep resources, the product ultimately failed. That experience reinforced a lesson that has guided me ever since: insight and talent aren’t enough. Without the right timing and ecosystem, even the best ideas stall. That lesson shapes how I invest at Lux Capital today. We back founders who aren’t just improving access to information, but expanding access to care itself—reducing friction, earning trust, and driving outcomes where the system is hardest to move. What’s different now is that the excuses are gone. Interoperability is real, with most U.S. hospitals exposing FHIR APIs and major EHRs supporting production integrations at scale. Clinical work is digitized enough for AI to live inside daily workflows. And the economic pressure is acute: administrative overhead still consumes 25–30% of spend, margins remain razor-thin, and clinicians generate nearly twice as much EHR data per patient as they did a decade ago. The result is a clear shift from pilots to purchases. We’re seeing rapid enterprise adoption of platforms like Abridge and OpenEvidence, justified by hard metrics—time returned to clinicians, capacity unlocked, and decisions made faster and better. Health systems aren’t adopting AI because it’s exciting; they’re adopting it because it’s one of the last levers left to keep care delivery viable. Healthcare is complex and regulated for good reason. But what stands out right now is alignment. Builders, clinicians, and operators aren’t chasing AI for its own sake—they want tools that work in the messiness of real care delivery. This feels like a genuine step function in both capability and relevance.

279

31 Comments

🧬 We talk about “health data” as if it’s one thing, but it’s really hundreds of incompatible languages trying (and failing) to talk to each other. Every layer speaks a different dialect: • EHRs: HL7 v2, CDA, FHIR • Claims: X12 837, UB-04, CMS-1500 • Labs: LOINC, SNOMED CT • Devices: DICOM, IEEE 11073 • Genomics: VCF, FASTQ, BAM Each was built for a single purpose, not interoperability. The result? 🚑 A patient’s data is scattered across 40+ systems, each with its own schema, timestamps, and access controls. But things are shifting. Newer models are moving beyond formats to: • Graph-based data structures • Semantic layers • Federated architectures These approaches preserve context, not just content, across systems. FHIR paved the road. But the next frontier is semantic interoperability. That’s not just data exchange; it’s data understanding. 🧠 The future of healthcare intelligence isn’t in collecting more data, it’s in connecting meaning. #HealthTech #DataInteroperability #FHIR #HealthcareAI #KnowledgeGraphs #SemanticWeb

147

15 Comments

🔥🔥 PANEL ALERT 🔥🔥 What gaps may AI fill with the recent policy impacts on Medicaid?  Coyote Ventures 2025 Health Equity Innovators Fall Gathering will feature an expert panel discussing the impact of AI in Healthcare with a particular focus on its implications for Medicaid and Medi-Cal. As we are past the mid-point mark of 2025, AI and OBBBA are reshaping U.S. health policy in profound ways, from expansive new federal and state guidelines to sharp debates around oversight and authority. Our discussion will dive into the themes of bias, equity, and accountability in AI, and what these evolving mandates mean for those developing, deploying, or relying on AI-driven solutions in healthcare. With Medicaid and Medi-Cal covering some of our most vulnerable populations, the stakes couldn’t be higher. I’m looking forward to an engaging conversation with Priyanka Vaidya, Pooja Mittal, and Stella Tran - exploring how we can ensure AI not only advances innovation but also builds trust, fairness, and better outcomes. Curious to hear your thoughts: can AI truly help reduce disparities in Medicaid care, or will it deepen them?

63

6 Comments