Happy Thursday Hospitalogists! Sword launched Pulse this week, an AI-enabled cardiometabolic care product, and I had a chance to chat with Kevin Wang, one of the most thoughtful dudes in healthcare who's running it, on launch day. Kevin is a longtime Hospitalogy reader and someone I've gotten to know through the health tech community - and I'm sure many of you have crossed paths with him. He walked me through Sword's model and thesis for cardiometabolic care - announcing Pulse this week - and I learned a lot. I'll be sharing those learnings and thoughts below - let's dive in! Also, follow Kevin if you aren't already. He will find a way to make me and my writing obsolete in T-minus 2 years, I'm fairly sure of it. |
Was this email forwarded to you? |
|
|
Going deeper on an interesting topic, theme, or trend |
An AI-Powered Bet on Cardiometabolic Care |
I want to start with Kevin's 'why' and his ownership of the product at Sword because I think it actually explains a lot about how Pulse was designed and why it feels different from the typical digital health cardiometabolic play.
Kevin has borderline high triglycerides (HIPAA breach!). He's had to fight his own doctors to get an advanced lipid panel - the classic "you're 30, your 10-year risk is low, we don't really need it" dismissal that anyone who's tried to be proactive about cardiovascular health (or any health, really) knows well. He's dealt with on-and-off essential hypertension that gets waved off at annual wellness visits. Meanwhile, his family history holds a history of cardiometabolic conditions.
So Kevin did what any frustrated health tech operator would do in 2025: he opened Claude, vibe-coded his own personal health assistant, and then walked into Sword CEO Virgílio Bento's (V) office who ironically had been thinking about the same thing. And that convergence of thinking is how Pulse came together. Pure vibes baby (kidding).
I bring this up not as a human interest aside but because I think it explains a design philosophy you can actually see in the product. Pulse is built around gradual, gentle behavior change over 4-12 months. It doesn't force people into strict keto diets or overwhelming protocols on day one. Kevin's frustration wasn't that cardiology care doesn't exist…rather that nobody helps you sustain the lifestyle changes that decrease downstream $. Behavior change and engagement is the holy grail of cost containment…and everything else that drives humanity, really. |
Sword's Growth Arc and Why You Should Care |
For those who haven't been paying close attention (you should be), let me give you Sword's quick trajectory. I made this graphic in June of 2024. It's now sorely, painfully (pun intended) outdated. In the first 2 months of 2026 alone, Sword has been on an absolute tear lately between acquiring Kaia Health, moving into mental health by launching Dawn, creating a comprehensive women's health platform, and now Pulse. |
After entering the U.S. in 2020 with Thrive, its digital MSK product, caught the COVID tailwind, and has been compounding since. We're talking 250 employees to over 1,300, nine figures in annual revenue, and what Kevin described as "triple-triple-double-double" growth. That's meaningful scale, not pitch deck math. Acquisitions have been channel-focused and deliberate - Kaia Health tacking on scale and lives, while other acquisitions or initiatives see Sword expanding into new territories like the UK, Germany, Canada, all while being bolstered by the talent back home in Portugal. Sword has sequenced its clinical expansion thoughtfully as it rebranded itself as an AI care company - intentionally comprehensive. MSK first (high urgency, high spend, lots of surgeries, clear ROI, natural employer demand), then Bloom for pelvic health, then mental health, and now cardiometabolic. Each vertical follows the same core playbook: Sword is a behavior change company, and behavior change methodology is transferable. If you can get someone to adhere to a PT regimen, you can get them to adhere to a nutrition plan. If you can build habits around movement, you can build habits around vitals monitoring. It's the same muscle applied to different problems. But then Sword builds operating leverage as patients across programs deal with similar things, and the beauty of the model plays out in operating leverage and streamlined patient care. |
Why Cardiometabolic and Why Now: State of the Market |
About 185M Americans have some form of cardiovascular disease, projected to hit 60% of the population by 2050. On the cost side, diabetes alone adds an average $7,500 in employer spend per person. Layer on hypertension, high cholesterol, prediabetes, which are conditions that cluster and compound, and you're staring at an enormous spend that self-insured employers or risk-bearing orgs are desperate to find meaningful solutions for. What ultimately gave Sword internal conviction to move in on heart health was their own data (and they also probably read some Hospitalogy notes about how red-hot the CV space is, womp womp). Sword had (has) been running a movement health program called Move designed for low-acuity pain and injury avoidance. When they actually looked at who was signing up, demographics showed ~ 3/4 of participants were obese or overweight, and more than half had hypertension, prediabetes, or high cholesterol. Move was attracting a cardiometabolic population organically without being designed for one. Move worked, too. 69% of physically inactive participants hit 150 minutes of weekly activity within 10 weeks. Kevin heard personal stories from members who told him their cardiologists took them off antihypertensives. OR people going from a size 22 to a size 16…prediabetes reversed. Kevin personally writes thank-you notes to members based on these stories. Has to make the job a bit easier when you see the impact you have. Anyway, since Move had such promising program results, Kevin & gang decided to take it a step further. |
AI as Care Team Member: Chapter Two of Sword and AI-Enabled Care Models |
Most health systems and digital health companies that claim to be "AI-powered" are really doing Chapter 1 stuff: ambient scribing, message recommendations for clinicians, documentation assistance. Important work, but it's low-hanging fruit. Chapter 1 is about making clinicians more efficient, and it's quickly table stakes. Chapter 2 (I prefer inning 2, honestly) - what Sword is building with Pulse - is AI as a direct member of the care team - agentic, automated. Pulse's interface serves as a three-way chat featuring the member, a human health specialist, and Phoenix, Sword's AI health specialist. Phoenix is a proactive conversational participant: - engaging members at 2 AM for shift workers,
- prompting meal logging at habitual times based on contextual memory,
- analyzing food photos and flagging concerns ("watch out for that sauce — half portion might be better, there's a lot of sodium"), and
- tagging the human clinician when escalation is needed.
We've seen a massive cultural shift in favor of consumer AI usage which increases comfortability with AI enabled products in healthcare and lowers onboarding friction dramatically, even for the demographics that digital health companies typically struggle to engage. Plus, making the AI a visible, named entity in the chat (rather than hiding it behind the scenes) creates transparency. Members know when they're talking to Phoenix and when they're talking to their human clinician which matters for trust and engagement. On safety, Sword has been putting real research into the public domain. Arbor, a conversational triage framework, dropped a few weeks ago. MindGuard, a mental health safety monitoring system, came before that. Both are open source. Publishing and open-sourcing AI safety research is a trust signal that I think matters, especially given how many companies are hand-waving their way through AI safety conversations with vague promises. Clinically, Sword currently staffs Doctors of Physical Therapy as the human specialists, a workforce Kevin argued is strategically underrecognized and describing DPTs as inherently behavior change practitioners since PT is fundamentally about building someone up to understand what they can do and feel confident doing it. Nurse practitioners are coming by year-end for medication management, prescribing, and titration, which is critical for a cardiometabolic product that plans to offer GLP-1 prescribing (but doesn't today). Connected devices round things out: a proprietary wearable with sensors including UV detection which I thought was pretty cool (indoor vs. outdoor awareness), blood pressure cuffs for hypertension, smart scales for prediabetes, and integrations with Apple Watch and Fitbit. Calendar and GPS integration enable context-aware nudges like "hey I noticed you're headed to the office, want to get off the bus one stop early?" and I love the nudges concept because nudges are known to be the kind of ambient, low-friction prompt that actually changes behavior over months, not just the first two weeks after someone downloads an app and then forgets about it. |
Outcomes-Based Pricing: Where Sword Gets Provocative |
Economics of these models are getting interesting. Most competitors price on an engaged PEPM basis with rates varying by disease state. Some - and Kevin was clearly frustrated by this - even risk-adjust their PEPM pricing and can't tell a buyer upfront what the cost will be. I think he called it "insane" if I remember his exact words. If you're a benefits manager trying to forecast costs and your vendor tells you they'll determine the price retroactively based on their own acuity assessment, that sounds shady even if it isn't intended to be. At best it lacks transparency. Sword's model across all products is an episodic case rate. Single price per member. 50% at activation (first clinically meaningful activity). 50% upon achieving clinically significant outcomes measured by the Patient Global Impression of Change (PGIC) — a well-validated, FDA-recommended 1-7 scale used extensively in clinical trials. Low-friction for the member, clinically defensible for payers, and backed by a 100% performance guarantee on claims-based ROI. Multiple clients have independently validated the ROI delivery. Plus…it's simple. Keep it simple, stupid. Sword could have stayed engagement-based and protected short-term revenue. Instead, Sword chose to use outcomes-based pricing as a forcing function: product teams literally cannot optimize for session counts or engagement vanity metrics when half of revenue depends on whether people actually get better. Very few have restructured their entire revenue model around something like this given business risk. Sword did, and I think that reveals something about how he thinks about building what Kevin called a "generational company." Obviously, this aligns philosophically with CMMI's ACCESS program, but the numbers, even for Sword, are tough. It'll be interesting to see whether they ultimately participate or not. |
Platform vs. Point Solution: Sword's Consolidation Thesis |
Zoom out from Pulse specifically and Sword's positioning gets really interesting from a strategic lens. Consider what they're pitching to a health plan: one contract covering 60-80% of employer healthcare spend. One CSM. One marketing plan. One dashboard. MSK, women's health, mental health, cardiometabolic — all under one roof. In a market where benefits managers are drowning in point solution fatigue, managing dozens of vendor relationships, and struggling to attribute ROI across fragmented programs, Sword's messaging is acutely pressing on multiple pain points (pun intended again - I'll stop now). And the platform economics actually hold up, which is the part that matters. Sword's central AI team builds shared architecture that individual products customize. Nutrition tracking built for Pulse gets leveraged by women's health for fertility and menopause use cases. Mental health's stress and sleep content flows into cardiometabolic programming. An internal repository called Academy houses educational content accessible across all products. Centralized data science, sales, and marketing teams create operating leverage that a standalone point solution — even a well-funded one — cannot replicate. Then there's Sword Intelligence, their separately branded AI health services arm. Care orchestration, referral management, medication adherence calls, AI care coordinators…Sword built these capabilities for itself, then realized health systems and plans would pay for them too. What compounds the moat over time is data and context. 800K patients treated. Over 10M AI care sessions since 2020. A proprietary AI research lab staffed through a Carnegie Mellon-University of Porto PhD pipeline (and low key an incredible asset and talent pipeline for AI research and development). In-house Vision AI. An NVIDIA partnership. Every new product vertical adds training data and behavioral signal that makes Phoenix smarter across the entire platform. McKinsey consultants drooling over the F L Y W H E E L effect at play in real time. V is reportedly deep in all of this. Kevin described him as doing "Founders Mode before Founders Mode became a thing" - in the weeds. Involved in chat design, product marketing videos, personally messaging the team about website memory card issues. For a nine-figure revenue company, that level of founder-in-the-details energy is a clear distinction. |
Pulse is the clearest signal Sword has sent that it sees itself as a platform - AI Care - company. Stack outcomes-based pricing on top of AI-as-care-team-member and shared infrastructure economics, and you've got something the point solution crowd can't easily replicate. I appreciated Kevin getting into the weeds of Sword Pulse with me, and it's a ton of fun to watch this company grow and develop new products. I hope they continue to succeed, thoughtfully infuse AI into their patient facing products, deliver incredible care and behavior change at scale, and reach the remaining 95% of folks in the process. Now they have to deliver. And I hope they do. Godspeed. |
Bonus: A Macro Question Worth Sitting With |
Kevin raised something at the end of our conversation that I haven't been able to stop thinking about, and I want to put it in front of you guys because I think it's one of the most important second-order questions in healthcare right now.
If companies like Sword succeed — if AI-enabled care delivery scales to the point where access really does become "as easy as flipping on the lights" — what happens to the healthcare employment base that's been propping up the U.S. economy since 2008?
We made a deliberate industrial policy choice after the financial crisis to invest in healthcare as a growth engine. Healthcare administrative roles employ millions of people and contribute meaningfully to GDP. Kevin referenced a Harvard Business Review study suggesting AI might actually create more work rather than eliminate jobs, and he framed it optimistically: more frontline clinical workers, fewer administrative positions, nurses returning to bedside care instead of drowning in documentation.
I want to believe that framing. But I also think "we'll redeploy admin workers into clinical roles" is a thing that sounds great at ViVE and is extremely challenging in actual labor markets. |
|
|
100 million US adults lack reliable access to primary care. The physician pipeline is short roughly 90,000 PCPs over the next decade. Our population is aging, chronic disease burden is rising, and the traditional fix — recruit more docs, build more clinics — doesn't pencil anymore. One health system calculated it would need to hire hundreds of new PCPs alongside physical clinics for all of them just to hit its patient acquisition growth targets. That's not a staffing challenge; it's a huge, unattainable capital requirement. Throwing humans at problems like the primary care crisis no longer works. It's time to enhance your primary care offering with a scalable AI-enabled solution. |
|
|
So I finally found a PCP in Dallas after a long hiatus and had to get some bloodwork done at Quest. You guys know how they always take your height and weight right? Well this chick did not care whatsoever if I had my shoes on or anything. So she took my height with my shoes, which happened to be some recent Hoka's I bought for walking. I'm proud to announce that my height in the Quest system is a smooth 5'10 and a half. My regular height? 5'9. |
|
|
Thanks for the read! Let me know what you thought by replying back to this email. — Blake |
| |
I'm building a community of leaders in strategy, finance, and ops at hospitals and health systems to help us connect, learn, and grow together. |
|
|
Get your brand in front of 66,800+ executives and healthcare decision-makers. |
Workweek Media Inc.
1023 Springdale Road, STE 9E
Austin, TX 78721 Want to ruin my day? Unsubscribe. |
|
|
|
No comments