A case for a moonshot in healthcare
Every era has a problem big enough to demand the very best of human ingenuity. In 1962, it was the original moonshot. Today, it’s healthcare.
When John F. Kennedy famously stood at Rice University, he called for the nation to come together to commit to a hard but solvable problem: the United States would put a man on the moon by the end of the decade. He recognized, like we do today, that the scientific means and the tools were both finally within reach, but the coordination required to put them into practice would be enormous. And if successful, the mission would lead to a step change in what society could do.
Today, we are looking at healthcare’s Apollo moment. Recent advances in AI, computer vision, and medicine have us on the brink of a generational shift. With it, we have everything we need to radically reimagine one of the most fundamental assumptions of our society: that healthcare is still a game of luck. Who you know, where you live, and what kind of resources you have should no longer determine your chances of survival and the quality of the life that follows.
It will not be easy, of course, but a new model of healthcare is possible and it deserves the very best of human ingenuity to deliver it.
The most important thing AI will do
At Sword, we’re building AI to heal billions and unlock humanity’s potential. The magnitude of our mission is reflective of our belief that artificial intelligence is capable of delivering a fundamentally new model of healthcare and that delivering it is the single most important thing that AI will do.
We’ve been developing the technical foundation to make this new model of care possible. That work has organized us around these technical focus areas, which remain the center of gravity for our work:
- We build proprietary frontier clinical intelligence, training large-scale models and building agents to deliver safe, measurable, and personalized clinical care.
- We change how people access and receive care by bringing together human care teams and AI to remove barriers to high-quality care. Our AI agents and AI Care solutions support every step of the patient journey, from triage and waitlists to prehab, rehab, and daily care across conditions.
- We publish research to advance the field through open scientific collaboration, focusing on pillars crucial to healthcare like evaluation and safety. Through this work, we establish industry frameworks to fairly and effectively measure clinical AI.
This is not something we can take on alone. It will take an effort within healthcare and across industries to reshape our healthcare system and deliver on the potential for AI in health. What we’ll share from here is how we’re thinking about this moment, the safety considerations required, and the future we’re working toward together with our clients and partners.
The future of healthcare is human
We believe the future of healthcare is clear: human clinicians and care teams supported and extended by purpose-built, clinical AI.
Healthcare has a long history of responding to workforce shortages by creating new roles; nurse practitioners and physician assistants emerged in the 1960s. Medical assistants and community health workers followed not long after. Each new role takes work that used to require a doctor and delegates it, freeing physicians to focus on the cases that require their full expertise. AI is the next chapter in that progression, but with a fundamental difference.
Every previous expansion of the healthcare workforce operated within the same structural limitation: care could only take place in a 1:1 setting. You book an appointment, see your specialist, and repeat. Nearly every problem in healthcare is a symptom of this underlying constraint. Our system is capped by the slots on a calendar, making care harder to reach. Waitlists grow, the privileged get seen sooner, and clinicians face burnout crises that only exacerbate the problem.
This system is inherently episodic; details get missed because what happens between appointments goes unrecorded. And as care gets harder to access, more fragmented, and more frustrating for patients, the threshold for seeking it gets higher. The window for prevention or early intervention gets smaller as people seek care later, raising costs for everyone involved.
AI unblocks access to world-class care
Even in the world’s most advanced healthcare systems, who receives quality care often comes down to luck and circumstance: who you know, where you live, and how much money you have to garner that 1:1 attention.
Incremental changes to this system will never solve the underlying structural constraint. Adding more human providers or introducing new clinician tools might temporarily relieve some pressure, but they ultimately reinforce the failure of the underlying model—one patient to one appointment. AI unlocks a new way of accessing and engaging with healthcare where the appointment is no longer the point of entry.
Of course, the natural question is, what would this feel like in practice?
Healthcare is, after all, the experience of being cared for by another person with their time, expertise, and empathy. Most of us are unwilling to trade that humanity for the efficiencies that technology can promise without some tangible proof that their care won’t get worse. In fact, it probably needs to get much better. Much like self-driving cars—we are unwilling to accept new risk on the road unless these cars are markedly safer than human drivers—the same goes for AI Care.
That’s why safety has been a core priority for us from the start. We believe the future of healthcare cannot run on general-purpose AI. It will take systems trained for the nuances of medicine, care delivery, and the complexity of modern healthcare organizations to meet this moment responsibly. A few principles and core philosophies guide our work in this domain.
- Clinical co-creation as methodology: we have clinicians embedded within our team, guiding our product development from the inside and informing the decisions we make for each area of care. They define what safe and effective care looks like based on evidence, clinical guidelines, and their personal experience. They also design our evaluation frameworks and validate that our AI behaves according to established professional standards and across diverse patient populations.
- Domain-specific safety: given the breadth of healthcare, even the best safety systems cannot flex for every case. Pregnancy guidance needs a specific set of guardrails, as does anxiety, injury treatment, etc. We build safety systems tailored to each clinical domain, grounded in the specific risk factors and standards of care that govern that particular practice.
- Evaluation grounded in clinical reality: Most AI benchmarks are divorced from clinical reality; they test knowledge through static questions and isolated responses. Healthcare doesn’t work that way. Recovery from chronic pain involves tracking progress over the course of weeks and months. Mental health support requires maintaining therapeutic rapport over many conversations. Our models are evaluated not only for their ability to provide the right answer in one moment, but also to deliver consistent care over time.
Even with these philosophical pillars in place, it is easier to see the risk of the unknown than to imagine the possibilities of it. Especially when many AI companies are committing their marketing budgets to building a distinct sense of doom.
We believe, perhaps paradoxically, that AI will make healthcare more human. It will free up physicians and clinicians to spend the time that patients deserve with the ones who need it. Patients will no longer have to repeat their story every time they have a new symptom or see a new specialist. Perhaps in the not so distant future, the idea of waiting weeks for an appointment or traveling miles to see a specialist will feel every bit as wrong as it is.
We are confident that healthcare will be markedly better tomorrow than it is today because we have imagined it in vivid detail. So let us paint a picture in a few scenarios of what AI Care can do today and where we envision it will go tomorrow.
Immediate care without the wait
Imagine an active-duty service member with low back pain. He has the kind of injury that’s easily treated with physical therapy, but waitlists for an appointment are 3-4 months. In the meantime, he’s still training and doing what it takes to fulfill his mission. By the time he’s seen, the problem has worsened considerably and what started as something treatable has instead become something surgical. With AI Care, he starts treatment immediately. His plan is personalized to his specific condition, adjusted by a Doctor of Physical Therapy as he progresses, and he can complete his training wherever he's stationed whenever his schedule allows. He avoids surgery because the problem is addressed upstream, before it compounds. He stays mission-ready, and the cost to the system is a fraction of what it would have been. Access to high-quality care at the right moment is also the cheapest form of care.
Personalized, proactive support
Picture a mom in her late forties. She’s been feeling exhausted for months, but chalks it up to a busy schedule and family life. She doesn’t see a doctor about it because nothing is wrong or clear enough to make time for an appointment. With AI Care, her wearable data shows a shift in her sleep patterns. Paired with self-reported fatigue and changes in her mood, the system recognizes what she hasn’t yet: the early signs of menopause. Her care team reaches out with a plan that spans nutrition, sleep habits, and a tailored exercise program—lifestyle adjustments that carry their own benefits and address the root causes of menopause fatigue. No appointment needed.
Continuous care between appointments
As a final example, consider a college student dealing with anxiety. She finds a therapist through her university's counseling center, but she’s limited to sessions once every two weeks with a fixed cap per semester. And for her, like most people, the hardest moments don't come during scheduled appointments; they come the night before an exam or in the middle of a crowded dining hall. With AI Care, a licensed psychologist guides her treatment, but she has 24/7 support. The system personalizes her care and carries a long-term clinical memory of what triggers her, which techniques have worked in the past, and how her patterns have shifted over time. It connects to her calendar and prompts her with the right exercise right before her exams. When she needs support in the moment, care is always there.
These stories are already possible today with AI Care.
This is the moonshot of our time
For the last decade, our focus has been excellence in a few, select areas of healthcare. We proved the model with MSK and physical therapy, then expanded to women’s health, mental health, and cardiometabolic health, all powered by a connected, clinical intelligence. In laying the technical groundwork to deliver it, we built agents and systems to efficiently triage and orchestrate care at population scale. Today, this fully verticalized system of care allows for expansion into virtually any area where the work is cognitive, rather than physical, and where access is artificially constrained by that underlying 1:1 model. Pediatrics, GI, primary care, cancer—each new area of care demands heavy investment, but the flywheel gets stronger as it expands. More data from more patient interactions trains better models, and better models deliver better outcomes for our patients and the people who care for them.
In that famous speech, President Kennedy called for mobilization not only within the government but across the entire nation. He asserted that this project would “serve to organize and measure the best of our energies and skills.” We believe reinventing healthcare with AI will do the same. There is no endeavor more worthy of our time, our investment, and the best of our abilities than delivering high quality healthcare to everyone who needs it. It will take each of us and all of us to make it happen.
And there is a role for everyone to play. If you’re responsible for the care of a population, your decisions today will determine whether the tools that are now available are finally put to use. If you are a clinician, your voice determines whether patients trust the systems being built for them. If you are a patient, your consent, or refusal, shapes what the next decade of healthcare is going to look like. Ask questions, do your research, imagine what better healthcare could finally look like. This is not a moment where the technology sector gets to build the future alone. It is a moment where everyone with a stake in healthcare and the human race has to decide whether we finally fix this system for good.
This is the moonshot of our time. Perhaps the most audacious project in modern healthcare and the most important thing that AI will ever do. On par with splitting the atom, mapping the human genome, and indeed, putting a man into space.
