Adam Russek-Sobol

Joined Bert Fernandez Fernandez on the HealthTech2Care Podcast to share one of our most exciting NIH-funded projects: We secured a Small Business Innovation Research (SBIR) grant to build a smell test data repository. The challenge: Loss of smell is linked to over 200 conditions from early Parkinson’s and Alzheimer’s to everyday safety risks like gas leaks. But, researchers lacked a domain-specific database and a way to easily query across datasets. So we built: → A centralized, NIH-funded repository for olfactory test data → A GenAI interface that lets researchers ask, “Show me patients with smell loss and cognitive decline” →A code-free experience—no R, no SAS, just natural language One researcher told us: “I can finally talk to the data, not write scripts.” This is how we make complex research accessible—and help fast-track early disease detection. 👉 Learn more about our work at Estenda Solutions: #GenAI #NIH #OlfactoryResearch #ExplainableAI #DigitalHealth #EstendaSolutions #HealthTech2Care

16

2 Comments

Most healthtech founders chase glamor metrics. Justin Quall chased the mess no one wanted to touch. He didn’t raise a seed round. He didn’t build a shiny dashboard. He didn’t start with a pitch deck. Instead, he moonlighted as an EMT. Read through 45,000 pages of discharge paperwork. And asked one deceptively simple question: Why are clinicians still choosing between documentation and patient care? That question became the foundation of Oler Health, a company quietly solving one of the most painful and overlooked problems in post-acute care. In our conversation on the Advancing Healthcare Innovation Show, Justin shares how Oler is helping skilled nursing facilities cut through bloated referral packets, surface the right clinical insights, and bring sanity back to intake workflows without removing the human from the loop. This isn’t another AI fairy tale. It’s a story about what happens when you build for the people who actually do the work. You probably know someone buried in documentation who needs to hear this. Feel free to send it their way...

62

6 Comments

Benchmarking the Future of Medical AI Agents: Really interesting paper introducing the first benchmark to evaluate how large language models function as medical agents inside EHRs. Traditional LLMs (like ChatGPT) act as chatbots providing answers to prompts. But "agents can proactively interpret high-level instructions, plan actions, interact with external systems, and iteratively refine their responses." That shift could help reduce clinician burnout, improve clinical care quality, and close workforce gaps. MedAgentBench tests 300 clinically derived tasks across 100 patient profiles (700K+ data points) in a FHIR-compliant virtual EHR. State-of-the-art models showed promise but with wide variation. Claude 3.5 Sonnet v2 topped the field at a 69.7% success rate. Impressive, but still far from reliable for clinical use where errors in medical records can have serious consequences. However, if current trends hold (as METR notes, AI’s ability to complete long tasks is doubling roughly every 7 months), today’s ~70% performance could reach clinically sound thresholds much sooner than expected. Just as MedAgentBench creates a virtual EHR to test AI agents in medicine, education needs its own suite of agentic benchmarks. Imagine an assessment agent that continuously measures student progress, an SIS agent (like an educational EHR) that integrates student records, a science of reading agent that adapts instruction to phonics and fluency needs, an HQIM agent that generates and sequences high-quality instructional materials, and a coherence agent that reviews all of this to ensure alignment with learning standards and the evidence base. Benchmarks could test whether these agents work together reliably—retrieving data, planning instruction, scaffolding practice, and adjusting in real time. Without such benchmarks, we risk deploying systems that look impressive in demos but fail to deliver meaningful learning gains. https://lnkd.in/gnzCwbi7

10

1 Comment

Can AI Safely Own the Rx Refill? 💊 AI enabling Telemedicine, here we go. Utah recently (January) became the first state to allow autonomous AI to handle routine prescription refills. The NY based startup behind it, Doctronic, just announced a $40M Series B round. I say this all the time; the bigger use cases for AI are for the backend use cases. Places where the clinician or patient may not realize it's there. Prescription refills are sitting somewhere in the middle. It needs data from previous visits, medical records etc. but also requires a visit or some medical interaction. This is using AI for both which is why it's considered autonomous. How the "Refill Bot" works: 🔹 Safety: It only manages refills for stable patients already under a clinician's care. 🔹 Medication guardrails: It operates on a strict formulary of 190 routine medications (no controlled substances). 🔹 Data Integrity: The system verifies the state prescription database and patient records before authorizing, rather than "taking the patient's word." This is standard for scheduled medications; for other meds, this is a difference. 🔹 A "Guardian" Layer: A secondary AI monitors the primary chat in real-time to detect medical emergencies. That monitor would be useful in itself. 🔹 The Human Loop: Every interaction is reviewed by physicians to ensure the "contained box" of automation stays closed. Earlier in the year there had been a version that was 'tricked' into giving a tripled amount of Oxycontin doses but the company states that wasn't the one used for patient interactions (and they are strictly avoiding scheduled med refills). While there is the fact that this type of services isn't reimbursed yet, I have an example of how a better refill program can still save costs for a health system program. Will tell you all about it next week. My takeaway: Innovation in 2026 is about the infrastructure and oversight. The FDA is still looking into safety concerns despite the guardrails the company has ensured. Clinician and patient trust cannot be overstated - especially if we are using something truly autonomous. #digitalhealth #AI #medicine #telehealth

25

13 Comments

I was so glad to participate in this insightful study with Kyra Gardner and Andy Wang as part of their program at Columbia led by Halle Tecco, MPH, MBA. I do believe AI will massively accelerate healthcare innovation, but right now it is stuck in a hype cycle and much of the early AI applications in healthcare just seem to be making the status quo more efficient rather than really transforming the system. We will get there....but there will be bumps and bruises along the way.

14

1 Comment

In this business, you either die a startup hero or live long enough to see your product outperformed by a Google Sheet macro. A new app popped up in the scripts I run daily today: "Autonephrine™ - Value Based Medical, Inc. [Patient-Provider Messaging] (Connection Hub)" While I certainly go a little hard in terms of checking out new APIs and interfaces that EHRs release, it's a bridge too far to look at every new app that pops up in EHR marketplaces. But the combination of the name (kinda cool riff on epinephrine) and the category (patient-provider messaging is a hot topic with AI tools today) piqued my interest enough to check it out more deeply. I was pretty floored to see that the app was a mixture of Google Forms and Sheets nicely bundled together. I did some digging, and it looks like a few anesthesiologists in California put it together for their own use. It's not clear which hospital or health system it's used with - the four providers are all California based but spread across geographies and hospitals ranging from University of California, San Francisco, Loma Linda University Medical Center, and Community Memorial Healthcare. This isn't intended to dunk on them or this app. Quite the opposite - it’s incredibly based to say, "I dislike my EHR and other tools so much I built a spreadsheet to use instead," and possibly the most successful clinician-driven innovation story you’ll hear all week. Beyond that, it's also worth noting that this app has already outperformed 95% of venture-backed startups by virtue of being in Epic Connection Hub. That means they are live with 1 or more health systems, which is a milestone most startups spend years chasing with far more capital, headcount, and swagger. Other things I like: - Very clear and specific value propositions communicated front and center on their web page (with real ROI) framed by real problems clinicians face: "Ever wondered how to do an M&M or QI without guilt or shame?" and "Is part of your group stipend at risk for first case start times?" - A transparent, self-service user manual that’s simple and approachable. - Straightforward technical setup documentation available publicly for IT staff. So props to Greg Stratmann, Adriana Bumb, Benjamin Kwitten, and anyone else using the tool. It's truly joyous to learn that somewhere out there, a Google Sheet macro is living its best life, outperforming entire product teams. And honestly? Respect.

261

40 Comments

700 healthtech startups. One underestimated state. Colorado’s not next. It’s now. ✅ 30+ standouts across AI, virtual care, and diagnostics ✅ Over $1.2B raised in the past year alone ✅ Deep activity from Seed to Series E, IPO, and acquisition SF and NYC still get the headlines. But Colorado is where early signals become category shapers. I mapped the breakout momentum: 21 unicorns and near-unicorns, across AI, virtual care, and diagnostics. It’s how I spot where real signals are forming, and which stories are worth following early. Building your edge? Check out my newsletter for more insights 🔗https://t2m.io/72LJ9BxF

34

1 Comment

Worlds First Open Source AI Based on Ethics! I am blown away with this open source and transparent platform. I met Eric Moore last year and when I saw what he was working on, I was immediately intrigued by his background and his vision. I have seen all the hard work and I have seen it come to life. Congrats Eric Moore! I believe this will be one of the standard AI platforms for industries where Ethical Reasoning, Compliance, and Accountability are Paramount. Look into it!

5

2 Comments

AI has officially moved into healthcare's waiting room. People are cancelling appointments with human nutritionists after getting tailored advice from AI platforms. Healthcare firms are benefitting as automation systems need less clinical oversight. It's happening faster than most of us expected. The regulatory reality check is real though. AI can assist, but it can't diagnose or prescribe. The winning approach blends AI efficiency with human expertise, keeping clinicians involved for complex decisions. For businesses, the opportunities are massive. AI reduces costs, improves access, and delivers personalised nutrition plans using everything from DNA profiles to real-time health data. But trust isn't built on clever algorithms alone. You need transparency about data sources, solid compliance with HIPAA and GDPR, and clear human oversight. This is where tools like n8n shine. They enable secure, flexible automation that meets the strictest healthcare standards whilst keeping everything connected and compliant. We've been working on exactly this kind of solution, blending AI with practical nutrition workflows. The results have been eye-opening, especially around how quickly personalised care can scale when you get the automation right. The conversation is moving quickly. Those building thoughtful AI solutions now, with proper ethical guardrails, will shape healthcare's future rather than just react to it. See what we've been up to in more detail, check out our case study in the comment below.

13

1 Comment

One of the most rewarding parts of building Alchemy is seeing our software used in community healthcare. I’ve spent the past 10 years building pharmacy software, so it’s especially meaningful when the tools we build extend beyond the walls of the pharmacy and into the community where care actually happens. Mobile testing is the foundation of the STD and Ryan White programs we serve. But the infrastructure supporting it has lagged behind. The workflows, the data capture, the reporting. All of it. This is not a clinical problem. It is not a funding problem. It is an infrastructure problem. Release #8: Mobile Testing Events is our answer to that. Shoutout to Jonathan Woong, Will Hollingsworth, Sara Shahdoost Moghadam, PharmD, MBA,BCACP,CDCES,AAHIVP who collaborated with clinic team members on the ground in Atlanta. We are grateful to Ronnie Bass and the entire Someone Cares team for your partnership in improving mobile testing in Atlanta. 👇 To learn more about this product release, see link to blog post in comments.

83

9 Comments

Today from Menlo Ventures: healthcare leads other industries in AI adoption! Previously a digital laggard, healthcare has flipped the script in AI, and providers lead the way. Startups won 85% of AI spend to date, but incumbents are fighting back. We just released "State of AI in Healthcare 2025" at HLTH USA with current data and insight from 700+ executives across providers, payers, and pharma. Read our research: https://lnkd.in/gcQKTXba Derek Xiao Johnny Hu Jean-Paul (J.P.) Sanday Croom Beatty

311

18 Comments

The most contrarian strategy in healthtech today? Not building with AI. Everyone is chasing the same LLM wrappers. The real, defensible moat is solving a non-scalable, human centered problem first. Then, and only then, apply technology to scale the solution.

4

1 Comment

Cleveland Clinic just partnered with Luminai, an AI startup that's tackling one of healthcare's most stubborn problems: administrative waste. The numbers tell the story: administrative work accounts for roughly a quarter of total U.S. healthcare spending. And despite years of digitization efforts, much of it still runs on faxes, manual data entry, and disconnected systems. Luminai's first use case? Referral management. Cleveland Clinic processes thousands of referrals daily—many still arriving by fax—each requiring manual review, data validation, and coordination across systems before care can begin. Early deployments are achieving over 80% automation rates for certain document types. What makes this partnership notable isn't just the technology—it's the strategic shift it represents. Cleveland Clinic isn't looking for point solutions anymore. They're building toward an AI-native operational layer that can work across multiple workflows. "We wanted to address workflows in a holistic way rather than by optimizing individual tasks," said Rohit Chandra, Cleveland Clinic's Chief Digital Officer. "This allows us to move beyond isolated tools and create a more cohesive, scalable approach." At Bobbin Health, we're watching this space closely. The shift from narrow AI tools to integrated platforms mirrors what we're seeing across healthcare—the question is no longer whether AI can help, but which solutions can actually scale beyond pilots and change how work gets done. Source: https://lnkd.in/efGcbti2

13

Hospitals and HealthTechs are trapped in a "Trust Tax". I was at ViVE last month. I heard the same fear from both sides. Hospitals: "How do I know my IP and sensitive patient data isn't being sucked into a vendor's black box?" HealthTechs: "How do I prove to my customers that I’m only taking the data I need?" Right now, the only "proof" we have is a BAA. But in 2026, a pinky promise isn't enough for a Board of Directors. When trust is low, the "Tax" is high. Sales cycles take 18 months because of security "paperwork" that doesn't actually stop a leak. How to fix: The Technical Airlock. If you're a hospital, it protects your "Digital Gold". If you're a HealthTech, it's your Competitive Advantage. The vendors who win in 2026 won't just say "Trust us." They’ll say "We literally can't see what we aren't supposed to. Here are the logs to prove it." Stop fighting over legal forms. Start building the vault together. #HealthTech #ViVE2026 #DataSovereignty #CISO #HealthIT #AIInHealthcare #CyberResilience #ZeroTrust #DigitalHealth

4

After lots of conversations with companies building in healthcare AI, one challenge keeps surfacing: the struggle to meaningfully integrate clinical expertise into a product. I mean, this isn't a new problem. It's existed long before AI became hot. Most companies I speak with operate on roughly 98% product expertise and 2% clinical expertise, if that. They're building sophisticated tools but often miss the nuances of real-world clinical workflows, leading to accuracy issues and adoption barriers. What if a company is 50% product expertise and 50% clinical expertise? What would clinicians be doing for these companies to transfer their clinical expertise into the final product? What value would we add and in what ways?

54

26 Comments

Here’s your recap of the last week's health IT news. 🗞️👇 🤝 𝗙𝘂𝗻𝗱𝗶𝗻𝗴, 𝗠𝗲𝗿𝗴𝗲𝗿𝘀, 𝗮𝗻𝗱 𝗔𝗰𝗾𝘂𝗶𝘀𝗶𝘁𝗶𝗼𝗻𝘀 • Healthcare AI operations platform Luminai raised a $38M Series B and announced a partnership with Cleveland Clinic to streamline referrals. • Maternal health company SimpliFed raised a $10M+ Series A and announced plans to offer virtual OB services to health system partners. • Behavioral health EHR vendor Qualifacts acquired controlled-medication dispensing software company MethodOne to integrate medication-assisted treatment. • Public safety and EMS software company ESO acquired healthcare analytics firm d2i. • Home-based care software company LivTech acquired home health EHR vendor Alora. 🚀 𝗣𝗿𝗼𝗱𝘂𝗰𝘁 𝗮𝗻𝗱 𝗣𝗮𝗿𝘁𝗻𝗲𝗿𝘀𝗵𝗶𝗽 𝗛𝗲𝗮𝗱𝗹𝗶𝗻𝗲𝘀 • Houston Methodist deployed Midstream Health’s AI financial intelligence platform across enterprise operations. • Reid Health deployed Epic’s Emmie to automate patient communication and administrative workflows. • Health systems AltaMed, Citizen Potawatomi Nation Health Services, Overlake Hospital Medical Center, Saint Francis Health System, and Valleywise Health became the first systems to connect to the Social Security Administration to exchange disability records electronically via Epic and TEFCA. • Premier Health Partners adopted clinical decision support tool UpToDate Expert AI system wide. • Southwest General Health Center deployed Oracle’s clinical AI agent to reduce physician documentation time. • Epic launched Health Alerts, which uses statistical models to detect when the rate of a health condition in a county is higher than expected. • Stedi launched new claims management tools for tracking and resolving payer claim issues. • Cedar expanded its AI platform to personalize patient billing communications and payment workflows. • Rune Labs launched an AI patient companion app to track Parkinson’s symptoms using data from its longitudinal dataset. • Talkiatry launched Talkiatry Connect to streamline psychiatric referrals and improve visibility for referring providers. • Zebra Technologies partnered with Aiva Health to enable hands-free nurse communication using voice and wearable devices • AdvancedMD launched an eMAR system to manage medication administration workflows in ambulatory settings

38

3 Comments

Utah just became the first state to legally let an autonomous AI system participate in prescription renewal decisions for chronic conditions, partnering with Doctronic through its AI regulatory sandbox. This pilot focuses on routine refills—an estimated ~80% of all medication activity—and aims to improve adherence, reduce avoidable ER visits, and cut into the more than $100B in preventable costs from medication non-compliance each year, while keeping clinicians at the center (“doctor, not device”). For anyone building in value-based care, pharmacy, or virtual care, this is a clear signal that regulated, outcomes‑tracked AI for high-volume, low-complexity workflows is moving from concept to reality—and that forward-looking states are willing to partner when safety, transparency, and clinical oversight are baked in from day one. I am wondering who bears the risk when it comes to medication errors or adverse effect? Doctronic? And hypothetically, lets say the outcome and cost effectiveness is far superior compared to the traditional model, like do we replace everyone with AI? https://lnkd.in/e_9kbB4V

20

14 Comments

𝗔𝗜 𝗜𝘀 𝗥𝗲𝗱𝗲𝗳𝗶𝗻𝗶𝗻𝗴 𝗥𝘂𝗿𝗮𝗹 𝗛𝗲𝗮𝗹𝘁𝗵𝗰𝗮𝗿𝗲 — 𝗡𝗼𝘁 𝗦𝗼𝗺𝗲𝗱𝗮𝘆. 𝗧𝗼𝗱𝗮𝘆. Rural healthcare providers aren’t just fighting disease. They’re fighting inefficiency, documentation overload, staffing gaps, and time. At the Center for Health Systems Innovation and OLSF Ventures, we believe the next wave of transformation isn’t about adding more people. It’s about amplifying the people already there. That’s why we’re excited about the work of companies like: Ambience – Using ambient AI to instantly document and code physician-patient encounters in real time. No scribes. No dictation. Just more time for care. Lumeris – Providing value-based care platforms and physician enablement services that optimize workflow, improve patient outcomes, and reduce unnecessary costs—especially important in resource-scarce rural markets. Percipio – Delivering advanced diagnostic support through AI-driven clinical decision tools, helping rural physicians confidently manage complex cases when specialty referrals aren’t an option. These aren’t abstract ideas or future-state visions. They’re live solutions—already deployed, already making a difference. Because in rural America, workflow is survival. If we want to close the health equity gap, we need to equip providers not just with compassion, but with the tools to do more with less. #RuralHealth #HealthcareInnovation #ArtificialIntelligence #Ambience #Lumeris #Percipio #CHSI #OLSFVentures #HealthEquity #Startups #DigitalHealth #AI #WorkforceEnablement #ValueBasedCare #Leadership #PrimaryCare

14

7 Comments