Giving Every Patient Their Own Chief of Staff

The United States has some of the finest doctors and medical institutions in the world. Our issue was never (primarily) the care; it's the system wrapped around the care.

It's the system that leaves patients waiting on hold. It's the system that forces a 74-year-old with three chronic conditions to figure out which specialist to see. It's the system that auto-denies claims, betting that patients won't appeal. Navigating this system is incredibly hard.

Nobody feels this more than the roughly 70 million Americans on Medicare and the family members who help them. Nearly a third of Medicare patients see five or more physicians and manage ten or more medications a year. Tens of millions of Americans provide unpaid care for a loved one, most while also working.

Consumers deserve to be CEOs of their own care. But a CEO without staff isn't empowered; they're overwhelmed. They need a team around them. We look for platforms that own longitudinal relationships with high-value populations and close the loop to care, unlocking outcomes-based measurement in the process.

Care navigation, done right, is the right wedge to give every patient a world-class chief of staff.

Why Care Navigation 1.0 Fell Short

Care navigation as a concept is not new. It's a $10B+ market that usually involves human navigators on the phone with patients as they move through the system, for example by answering questions about benefits or helping patients find the right specialist.

Programs that focus on payers and self-insured employers often see low patient engagement and, at minimum, the perception of misaligned incentives. Fundamentally, these offerings are often reactive or episodic, resulting in low retention and limited impact.

Due to the labor-heavy nature of these models, it is difficult to economically justify long-term patient engagement even if you can earn it. And as with many areas within healthcare, care navigation 1.0 provides limited visibility into outcomes, driving little by way of efficiency gains and therefore a weak data moat.

What Care Navigation 2.0 Looks Like

Standalone, navigation needs are episodic and reimbursed based on time vs. value. However, the best founders in this space realize that care navigation is the right wedge for a longitudinal patient relationship. Over time, they earn the right to be a patient's first stop when engaging with their health.

From there, the winners will look different from 1.0 in four ways:

1. DTC first, but not only. Care navigation 2.0 starts direct-to-consumer, proving strong engagement and patient trust, which can then be scaled through payer and provider partnerships. Acting as the patient's advocate, not the payer's or employer's, is what earns the relationship.
2. AI-native, with targeted human labor. AI collapses the cost of delivery, so the human advocate's time goes where it matters most.
3. A suite of connected services, including care delivery. Navigation surfaces needs, providing a clear product roadmap for growth.
4. Outcomes from day one. Every interaction compounds into a longitudinal record that can prove what works, and that record becomes the moat. AI also enables a step-function change in fraud, waste, and abuse detection.

The Payment Problem for Patient-Facing AI

For seniors, coverage is particularly critical: in 2024, half of people on Medicare lived on less than $43k per year. Fortunately, Medicare introduced new billing codes that same year to reimburse care navigation. These non-clinical services address health-related social needs and patient guidance through a serious physical or mental illness. CMS coverage unlocked a once-in-a-decade new market, finally creating an efficient way to serve seniors amid ongoing health system complexity and high rates of chronic conditions.

The most significant barrier to widespread clinical AI adoption is payment models. Healthcare providers are paid for procedures, visits, and time spent with patients. They don't get paid for AI-generated diagnosis, monitoring, or preventive interventions.

Even with insurance coverage, these rates are time-based and tied to human labor. For covered navigation, an AI agent cannot independently bill. Any AI must be deployed under the supervision of clinicians or navigators to fit the current codes.

Value-based care models solve this by paying for outcomes rather than services, but only 30-45% of US healthcare operates under value-based contracts. New programs, including CMS's ACCESS model, seek to test outcomes-based reimbursement, but for many providers the current fee-for-service models still make more economic sense, at least today. But if such tools significantly improve efficiency, CMS could eventually consider new models, for example enabling some automation within the service, akin to how Remote Patient Monitoring allows automated data collection with clinician oversight.

How Medicare Leads

We believe the right way to change the system is to work around it where reasonable (non-reimbursable actions that patients will pay for), with it where sensible (existing CPT codes), and redefine it where possible (helping define new care delivery and coverage models).

Medicare plays a leading role in supporting healthcare innovation. Its size gives it significant pricing power, and commercial insurers often follow Medicare's lead on coverage. Stable enrollment bases with high utilization create the right conditions for value-based care. Further, Medicare has a billion dollar annual innovation budget, managed through CMMI. Privately run Medicare Advantage plans compete via innovation in benefits, care models, and risk-based contracting.

Despite strong coverage, MA plans are under significant margin compression driven by risk score reform, post-COVID utilization trends, and market maturity. Medicare accounts for 21% of US health spending, and as 10,000+ Americans age into Medicare every day, managing costs becomes increasingly paramount.

These are the right conditions for change. Regulation is forcing a shift from documenting risk to actively managing outcomes, with V28 reducing passive risk-score optimization and value-based contracts requiring longitudinal intervention. That makes care navigation the core orchestration layer of value-based care.

Earning the Right to Grow

With a huge unlock in demand comes a responsibility to guard against fraud, waste, and abuse. These codes are not technically capped in the number of hours a Medicare beneficiary can receive per month, but CMS sets volume thresholds, known as Medically Unlikely Edits, above which billing requires additional justification. The companies that last will treat those guidelines as a floor for integrity, not a target to optimize around, and use AI to make sure patients get the most out of every hour of their navigator's time.

We look for founders who can demonstrate clinical value and safety through rigorous validation, and define those benchmarks for others. AI collapses the cost of delivery so much that it can be worth taking a short-term hit to revenue. Especially in areas where outcomes can be clearly measured, the revenue models will catch up.

Why This Compounds

AI will make care dramatically cheaper to deliver. The question is who keeps the value. In most industries, cheaper production gets competed away. Healthcare is different. State licensure, payer contracts, access to patient data, presence in local communities, and above all trust are hard-won and slow to replicate, and they let the value of better technology accrue to the provider. That's why we believe some of the most durable compounders of the AI era will be built in healthcare.

As we invest across consumer health, wealth, and happiness, ensuring every patient has a world-class chief of staff sits at the intersection of all three.