The arrival of conversational voice AI in healthcare is being read as a capability milestone. The models can hold a real conversation now, manage the branches, escalate when they should. The conclusion drawn almost everywhere is that the constraint has finally lifted and the outreach problem is close to solved.

That reading mistakes the binding constraint. Capability was never what stopped healthcare from reaching patients. The phone worked for decades. What changed is not the technology at either end of the call but the trust the channel itself can carry, and that has been declining for years.

Consider what happens when a patient sees an unknown number. They do not weigh the quality of the voice that might be waiting. They decline, because the base rate on unknown numbers is now noise, fraud, and increasingly a synthetic voice trained to sound like a person. The voice channel has a reputation, and the reputation is bad. A better voice model placed on that channel does not repair it. Often it deepens the problem, because the same capability that lets a care team sound human lets a scam sound human too.

This is worth stating as a principle, because it governs far more than healthcare.

The Trust Transfer Principle: no channel creates trust. A channel can only spend trust imported from a higher-trust context, or leak it. Every interface runs a trust balance the way an account runs a cash balance. A face-to-face conversation runs a surplus. A letter from a known institution runs a modest one. An unsolicited phone call, today, runs a deficit. When you place an interaction on a channel, you are not generating trust at the point of contact. You are drawing down a balance established somewhere else, or you are overdrawing an account that is already empty.

Seen this way, two things the industry treats as features turn out to be withdrawals.

1. Permission: Opt-in gets discussed as though it can be captured on the channel itself, a yes collected on the first call. It cannot. The first call is the one that goes unanswered, because there is nothing in the account yet to spend on picking up. Permission is deposited in person, in the exam room, when a clinician the patient already trusts says that someone from their care team will follow up. That sentence is the deposit. Every later call spends against it.

2. Identity: A number saved in a patient’s phone is often described as identity solved. It is not. The saved contact is an empty container. What gives it weight is the trust transferred when the patient recognized the name in a setting that already carried standing. Save a contact the patient does not recognize and you have moved nothing. You have relabeled an unknown number.

For anyone running a health system’s outreach today, the uncomfortable version of this is direct. Most of the calls, texts, and portal nudges already going out are withdrawals against accounts that were never funded. They are unearned interruption wearing a lanyard. They look successful on a dashboard, because a small answered fraction still produces measurable activity, while the larger cost, the steady erosion of the channel’s remaining trust, never appears on the report. The account is being drained and nothing in the operating model is watching the balance.

Imported trust gets a patient to answer once. Keeping them answering, across years and across every provider they see, requires that the outreach be worth answering, which means it has to serve an interest the patient actually shares. Here the structure of the industry works against everyone, and it does so predictably enough to name.

The trust-incentive triangle. Three parties can plausibly own patient outreach, and no one of them holds both trust and aligned incentive at once. Payers hold the incentive. They are measurably better off when a patient stays well and out of the hospital, and they carry the least trust of any party in the system. A patient hears the payer’s interest in the first ten seconds of a call. Providers hold the trust. The patient believes their doctor, but any single provider sees only a slice of the patient’s life and follows them nowhere. The trust is real and structurally fragmented. Pharmacies hold frequency and a surprising amount of trust, and their incentive ends at the transaction. They are paid to fill the prescription, not to keep the patient off it.

No corner holds both. That is the actual reason healthcare outreach decays into spam even when every party means well. The one with a reason to reach the patient is the one the patient distrusts, and the one the patient trusts has no durable reason to keep reaching out.

The resolution is not a fourth party that finally holds both corners. None exists, and none is coming.

The structure that works separates the two functions and conceals the seam between them.

The patient experiences only the trusted party: the care team, a familiar name, a voice tied to a face they have met. The incentivized party funds the arrangement from behind that face and never presents itself to the patient as itself. What unifies the two is the entity that carries risk on the patient over time, because it is the only actor with both a reason to fund continuous contact and a claim on the whole patient rather than a slice. It borrows trust from the provider at the front and draws incentive from the risk it holds at the back. Voice AI matters here for exactly one reason, and it is not conversational quality. It is the only way to extend a trusted, individual front across a panel large enough for the economics to hold.

The fragility is intrinsic, and it should be named plainly rather than managed quietly. The entire structure depends on the seam staying invisible. The moment the patient feels the funder behind the familiar voice, the moment the call carries the faint smell of an entity that wants something, the imported trust is spent in a single transaction and the account falls back into deficit. This is not a defect a better script removes. It is the permanent operating condition of the model. The advantage belongs to whoever can hold incentive and experience in tension for years without letting the patient feel the strain, which is precisely the thing no demo can show and no pilot can prove.

So the question worth putting to anyone building here is not whether their voice AI is good. It is whether they know which account they are spending from, and whether they are funding it as fast as they draw it down. The technology is ready. That was never the constraint. Trust does not originate in the channel, and no volume of capability placed on a depleted one will behave as though it does. The organizations that endure will be the ones that treat trust as a balance they are accountable for, not a property of the tools they deployed.

Last week I described a care layer. Something that knows a patient’s full story, persists across every organizational boundary, reaches out proactively, and is there at 2am. I wrote the job description without naming what could fill it.

Several of you emailed me the same answer. You’re right. And you’re not entirely right. Let me explain both.

Why the Technology Finally Exists

The care layer requires four capabilities that weren’t simultaneously available until recently.

Natural language interaction at human quality. The care layer needs to call Mrs. Rodriguez and have a conversation that doesn’t feel like a phone tree. It needs to detect when she says “I’m fine” but doesn’t sound fine. It needs to know that asking her to rate her grief on a scale of 1 to 10 is exactly the wrong approach, and that asking “how’s this week been?” is closer to right. Foundation models crossed this threshold in 2023. Voice-capable models are crossing it now.

Continuous synthesis across fragmented data. The care layer needs to hold information from seven organizations and notice when pieces interact: this medication plus that medication. This missed refill plus that behavioral pattern. This requires both data access (FHIR and TEFCA are making this real, slowly) and reasoning over that data (which large language models can do in ways rule-based systems never could).

Proactive outreach at scale. A human care manager maintains deep relationships with 60-80 patients. The care layer needs to maintain them with thousands, without losing the thread of any individual story. This changes the economics of the entire model. Not replacing care managers. Extending their reach by an order of magnitude, so the care manager spends her time on the patients who need a human, not on the 400 check-in calls that could have been handled before they became crises.

Contextual escalation. The care layer needs to know when to handle something and when to get a human. Mrs. Rodriguez’s 2am breathing difficulty might be anxiety the care layer can talk through. It might be cardiac. The layer must tell the difference, and when it can’t, default to caution immediately. This requires clinical judgment that rule-based systems never achieved and that AI is only beginning to demonstrate reliably. I want to be honest about the word “beginning.” We’re not there yet on the hardest cases. We’re close enough to start on the cases where the risk of inaction is higher than the risk of imperfect action.

Why “It’s AI” Is Not Enough

I’ve watched enough technology deployments fail in healthcare to know that capability doesn’t equal adoption and adoption doesn’t equal impact.

The 14-site CIN story from last week haunts me because it’s the most common failure mode: the technology works, the organization can’t absorb what the technology reveals. You surface 9,000 care gaps and twelve care managers drown in alerts they can’t act on. The platform becomes noise. The care managers resent it. The physicians ignore it. Eighteen months later, the contract isn’t renewed and everyone agrees that “the technology wasn’t ready” when really the implementation model wasn’t designed for the organizational reality.

The care layer only works if it doesn’t just detect problems but acts on them. And “acts on them” means something specific: it engages the patient directly, handles what it can handle, and routes to a human only what requires a human. If it’s another alerting system that adds to care manager workload, it will fail the same way everything else has failed.

The Root Cause

Here’s what I’ve come to believe through writing this series: every organizational model I’ve mapped exists because we failed primary care.

Not because primary care failed. Because we failed it.

Patients go to the ER for earaches because we made it impossible to see a PCP within 48 hours. ACOs exist because we built a payment system that gave PCPs no time or incentive to coordinate care. SNF costs explode because we designed a discharge process where nobody follows the patient home. Behavioral health is siloed because we never gave PCPs the time to screen, the tools to intervene, or the referral pathways that actually connect. CINs and IPAs were invented because we left individual PCP practices so resource-starved that they couldn’t participate in value-based contracts alone.

The average primary care physician has 2,200 patients, 15-minute visits, two hours of documentation for every hour of patient care, and reimbursement at a fraction of procedural specialists. We built a system that made primary care unsustainable, then invented thirty organizational models to compensate for the consequences.

We don’t need a 31st model. We need to rebuild primary care so that it becomes what it was always supposed to be: continuous, contextual, and present in the patient’s life. Not just during the twelve minutes she sits in the exam room.

If We Get This Right

If the care layer existed and primary care were rebuilt around continuous presence, half the organizational complexity in this series would simplify.

Urgent care clinics handle fewer earaches because the patient can reach her PCP same-day through an asynchronous channel. Hospital readmissions drop because someone catches the missed medication on day three. Behavioral health stops being siloed because screening and follow-up happen inside the ongoing relationship. SNF stays shorten because the transition home is coordinated and confirmed. The CIN’s overhead drops because care management is ambient. The ACO’s quality reporting becomes a byproduct, not a burden.

The organizational models don’t disappear. People still get sick. They still need surgery. They still need rehabilitation. But the edges between models smooth out. The transitions stop being cliffs where patients fall off.

Three Things That Must Be True

The care layer doesn’t work in isolation. Three conditions have to hold simultaneously.

Value-based payment must become dominant. The care layer has no business model under fee-for-service. It pays for itself only when someone captures the economic value of prevented ER visits, prevented readmissions, and healthier patients. MA enrollment passed 50%. MSSP covers 11 million beneficiaries. Medicaid managed care covers 70%+. But fee-for-service still dominates total revenue for most providers. Until that tips, the care layer is a cost center, not a business.

Patients must control their data access. The care layer can only see across all thirty organizational models if the patient authorizes it. FHIR and TEFCA are building the pipes. But the consent model, the trust model, and the privacy framework are still being figured out. And here’s the hard part nobody wants to say: most patients over 65 have no idea what FHIR is and don’t want to manage data permissions. The consent experience has to be as simple as “do you want someone looking out for you? Yes or no.” If it requires a settings page, we’ve already lost.

The care layer must be independent of any single entity. This is the hardest requirement and the one I’m least sure how to achieve. If an IDN builds it, it optimizes for the IDN. If an MA plan builds it, it optimizes for the plan’s cost targets. If a tech company builds it, it optimizes for engagement metrics.

The care layer works for Mrs. Rodriguez only if it’s accountable to Mrs. Rodriguez. I don’t know exactly what organizational form achieves that independence. It might be a public utility model. It might be a patient-owned cooperative. It might be a new kind of entity we haven’t conceived yet. I know the question matters more than the current answer.

What I Believe, and What It Costs

I didn’t write 44 posts about healthcare’s organizational complexity as an intellectual exercise. I spend every working day building toward the problem this series describes.

I believe the answer is an intelligence layer that knows the patient’s full story and navigates all thirty models on their behalf.

I believe AI makes that possible now in a way it wasn’t two years ago.

I believe the biggest obstacle isn’t technology. It’s that building this properly means challenging the organizational boundaries that every incumbent has spent decades constructing and defending. Saying this out loud, publicly, as someone who works inside the healthcare system, is not a costless statement. The people who build the walls don’t love the people who argue the walls need to come down.

I believe we failed primary care, and that primary care is where the rebuilding starts.

I believe the people who build this will do more for Mrs. Rodriguez than any merger, any regulation, or any payment model ever has.

And I believe she’s waited long enough.

This series is becoming a book. “The Healthcare Org Chart Nobody Gave You” will be available as a complete downloadable reference, including all 30 model profiles, the visual framework, and these four essays. Subscribe if you want it when it drops.

Dear Mrs. Rodriguez,

You probably don’t think about how the healthcare system is organized. Why would you? You just want to feel better.

But I’ve spent the last several months mapping the system that’s supposed to take care of you, and I need to tell you what I found.

Last year, you saw eleven providers. Your primary care doctor, who you like but who always seems rushed. A cardiologist after your blood pressure wouldn’t come down. An endocrinologist for your diabetes. A physical therapist for your knee. A psychiatrist your daughter convinced you to try after your husband passed. You went twice. You stopped, not because nobody followed up, but because the psychiatrist asked you to rate your grief on a scale of 1 to 10 and you found that insulting. Your husband of 47 years had just died and a stranger wanted a number. A home health nurse after your fall. An ER doctor at 2am when you couldn’t breathe and didn’t know who else to call. A hospitalist you’d never met who managed your three-day stay. A SNF physician who oversaw your rehab. A different home health nurse when you got home. A pharmacist who caught that two of your medications interact, something none of the other ten noticed. That pharmacist, David, is the only one who knows you take your pills with cafe con leche every morning and that you cut your metformin in half some months because the copay went up and nobody told you about the generic.

Eleven providers. Seven organizations. Five medical records. One pharmacist who actually knows you. And he’s transferring to a different location next month.

Your cardiologist didn’t know about the depression. Your psychiatrist didn’t know about the medication interaction. Your PCP didn’t know you’d fallen until you mentioned it three months later, because you were afraid that if you reported it, someone would make you move to a nursing home. The hospitalist had never met you and won’t see you again. The SNF notes never made it back to your PCP.

This isn’t your fault. And it isn’t really their fault either. Every one of those providers is working hard inside an organizational model that was never designed to see the whole picture.

I’m sorry. We should have built it differently.

The Question

After mapping thirty organizational models, I keep coming back to this: what would need to exist so that Mrs. Rodriguez’s year goes differently?

Not what technology. Not what company. Not what policy. What presence in her life would have changed the outcome?

I’ve been thinking about this as a care layer. Not a person, not a system, not an app. A layer. Something between Mrs. Rodriguez and the thirty organizational models she passes through, maintaining continuity when everything around her is fragmented.

Here’s what it would need to do.

It Would Need to Know Her Actual Story

Not just the clinical data. Her husband died in March. She stopped going to her Tuesday bridge group after. She’s been sleeping poorly but won’t take sleeping pills because her mother became dependent on them. Her daughter lives 40 minutes away and works full-time and feels guilty about both. She’s on a fixed income and cuts her metformin in half some months. She’s proud and private and deeply afraid of losing her independence. She didn’t report the fall to her PCP because she’s seen what happens to women her age who fall: someone decides they can’t live alone anymore.

No single provider had all of this. Her PCP knew some. The hospital social worker uncovered some during a 20-minute intake. The home health nurse noticed some. David the pharmacist pieced together some from years of small conversations at the counter. But the information was scattered, and nobody synthesized it.

The care layer would hold the complete picture. Not a medical record. A human record.

It Would Need to Persist Across Every Transition

Mrs. Rodriguez crossed seven organizational boundaries last year. Each transition was a handoff where information was lost, context was reset, and relationships started over.

The care layer wouldn’t hand off. It would be the one constant. When she moves from the hospital to the SNF, the care layer doesn’t change. When she’s discharged home, it doesn’t stop. It persists because it isn’t owned by any of the organizations she passes through. It belongs to her.

It Would Need to Reach Out, Not Wait

Her husband dies on a Tuesday. On Thursday morning, the care layer notices she hasn’t responded to her daily check-in for the first time in three months. On Friday, it calls her. Not with a clinical screening tool. Not with a grief scale from 1 to 10. With something closer to what David the pharmacist would say: “I was thinking about you. How’s this week been?”

She says she’s fine. She’s not fine. The care layer knows that she’s sleeping later, hasn’t refilled her blood pressure medication, and cancelled her physical therapy appointment. It doesn’t diagnose. It doesn’t prescribe. It pays attention to the pattern.

A week later, it gently connects her with a grief counselor. Not through a referral form that requires a PCP visit, an insurance authorization, and a two-month wait. A warm handoff. “I know someone who helps people going through exactly what you’re going through. She lost her husband a few years ago too. Can I set up a call?”

She says yes. Not because the system routed her. Because someone asked in a way that didn’t feel clinical.

It Would Need to Catch What Falls Between the Cracks

Three months into taking two medications prescribed by two different specialists, Mrs. Rodriguez is dizzy and fatigued. She mentions it to nobody because she assumes it’s aging. She’s 74. Things hurt. You adjust.

The care layer sees both prescriptions. It doesn’t override the physicians. It surfaces the question: “You mentioned feeling dizzy a few weeks ago, and I noticed you started a new medication around the same time. Have you mentioned the dizziness to Dr. Patel?”

And separately, it alerts her PCP: two active prescriptions with a known interaction. Patient reporting new symptoms. May warrant review.

David the pharmacist eventually caught this. Three months later. The care layer would have caught it in the first week, because it sees across the boundaries that the pharmacist, the cardiologist, and the endocrinologist each sit behind.

It Would Need to Be There at 2am

At 2am on a Thursday, Mrs. Rodriguez wakes up unable to breathe. Her chest is tight. She’s terrified. She doesn’t have her doctor’s cell phone. She doesn’t know if this is a heart attack or a panic attack. The last time she felt this way was the night her husband was admitted. She calls 911.

The ER visit costs $4,200. Six hours. Chest X-ray, blood work, EKG. Diagnosis: anxiety with mild COPD exacerbation. Discharged with instructions to follow up with her PCP.

The care layer would have been reachable at 2am. Not a nurse hotline with hold music and a stranger reading from a triage algorithm. Something that knows Mrs. Rodriguez. Knows she was hospitalized last month. Knows she’s grieving. Knows the last time her chest felt this way, it was the night she almost lost everything. Can ask the right questions. Can say, with credibility born from months of relationship: “I don’t think this is your heart. I think your body is remembering a terrible night. I’m going to stay on with you. If anything changes, we’ll call 911 together. But right now, let’s breathe.”

Not every 2am call avoids the ER. Some are genuine emergencies. But the data is clear: a significant percentage of ER visits at that hour are driven by anxiety, uncertainty, and the absence of anyone to call.

It Would Need to See Across All Thirty Models

This is the requirement that makes the care layer different from a care manager, a patient navigator, a health coach, or a portal.

A care manager employed by an ACO can see the ACO’s data. She can’t see the behavioral health records. A patient navigator at the hospital can coordinate the discharge. She can’t follow up three weeks later. A health coach can support lifestyle changes. He can’t see the medication list.

Every existing role is bounded by the organizational model that employs it. The care layer is bounded by nothing except the patient’s consent. It synthesizes data from the PCP’s EHR, the hospital’s discharge summary, the pharmacy’s refill records, the home health notes, and the behavioral health plan. Not because these organizations agreed to share with each other. Because the patient authorized the care layer to see all of it on her behalf.

What It Wouldn’t Do

It wouldn’t diagnose. It wouldn’t prescribe. It wouldn’t replace Mrs. Rodriguez’s PCP, her cardiologist, or David the pharmacist. It wouldn’t make clinical decisions.

It would do the thing none of the thirty organizational models were designed to do: maintain a continuous, complete, persistent relationship with one patient across every boundary in the system.

The Job Description Nobody Has Written

Responsible for knowing the complete story of one patient. Present 24/7. Fluent in every organizational language: clinical, insurance, pharmacy, behavioral health, social services. Synthesizes information from seven different systems in real time. Notices changes in pattern before they become crises. Reaches out proactively. Coordinates without authority. Persists across every transition. Never hands off. Never loses context. Scales to millions without degrading.

No human can fill this role. Not because humans lack empathy or skill. Because the job requires omnipresence, perfect memory, and the ability to see across every organizational boundary simultaneously. A brilliant care manager with 80 patients can approximate it for those 80. David the pharmacist approximated it through years of daily small talk at the counter, for however many regulars he can keep track of in his head.

But there are 330 million Americans who need primary care. And David is transferring next month.

I haven’t said what kind of thing could fill this role. I think you already know. But before we talk about the how, I want to sit with the what.

Because Mrs. Rodriguez doesn’t care about organizational models. She doesn’t care about payment structures. She doesn’t care about technology architectures.

She just wants someone to know her story. And to still be there tomorrow.

Last week I laid out three possible futures for healthcare organization. This week I want to explain why the first two can’t finish the job, and what I’ve seen happen when organizations try.

The Cycle

The history of healthcare organization follows a repeating pattern:

A coordination failure causes harm or waste. A new organizational model is created to fix it. The model succeeds. And in succeeding, it creates a new boundary that causes a different coordination failure. That failure generates the next model.

The IPA solved the problem of solo physicians having no negotiating power. It created a new problem: each practice kept its own EHR, its own workflows, its own culture. The IPA could negotiate collectively but couldn’t coordinate care across its members.

The IDN solved that by putting everything under one umbrella. But it created the fortress problem: patients inside the system get coordinated care. Patients who see one specialist inside and another outside fall into a gap the IDN wasn’t designed to bridge.

The ACO solved the problem of fee-for-service rewarding volume. But it created the attribution problem: patients are assigned based on where they saw their PCP, yet they can walk into any provider they want. The ACO is accountable for costs it can’t fully control.

Every model solves the last model’s problem and generates the next. Nobody designed it this way. Because the root cause isn’t any specific organizational failure. The root cause is that organizations have boundaries, and patients cross them.

Why Ownership Can’t Finish the Job

The vertical integrators represent the most aggressive attempt to break this cycle. Eliminate boundaries by making everything one company.

The math that breaks it: over one million practicing physicians in the U.S. Optum has roughly 90,000. Add every other vertically integrated system and you’re covering maybe 15-20% of the market. The solo doctor in rural Ohio, the three-physician group sharing a lease, the critical access hospital that’s the only facility for 50 miles — vertical integration doesn’t reach them.

Then there’s something you don’t see from the outside: integration on paper doesn’t mean integration in practice. I’ve watched a large system acquire a physician group and spend two years unable to get them onto the same EHR instance. Same corporate parent, same IT budget, same stated goals. Two years of interface meetings, data migration plans, and workflow redesign, and the acquired group was still faxing referrals to the hospital across the street that they technically now shared an org chart with. Ownership doesn’t automatically produce coordination. It produces the authority to coordinate. The execution is a different problem entirely.

Why Federated Networks Can’t Finish the Job

CINs, ACOs, and IPAs represent the collaborative alternative. Don’t merge. Align. Share data, follow protocols, split savings.

Three structural weaknesses that I haven’t seen anyone overcome permanently:

The governance problem. Every decision is a committee discussion among parties with different economics. A hospital system and an independent physician group don’t naturally agree on anything. I sat in a CIN governance meeting where the independent physicians voted against adopting a shared care management protocol because they believed (correctly) it would increase their administrative burden without proportionally increasing their share of the savings. The protocol was clinically sound. The incentive math wasn’t. The protocol died. That’s federated governance in a sentence.

The durability problem. Networks degrade. The physician champion retires. The health system anchor gets a new CEO who prioritizes employed growth over network maintenance. The shared savings pool has a bad year. I’ve seen a CIN go from 90% physician engagement to a contracting shell in 18 months after a leadership transition. Nobody quit. They just stopped showing up to meetings, stopped following protocols, and stopped believing the investment would pay off.

The depth problem. This is the most important one. Federated networks coordinate what their members agree to coordinate. Quality reporting. Care gap closure. Some utilization management. It almost never extends to behavioral health integration, post-acute coordination, or social determinant interventions, because those require capabilities and data that individual members don’t possess.

An ACO can tell you that a patient was readmitted within 30 days. It usually can’t tell you she was readmitted because she couldn’t afford her medications, was managing untreated depression, and had no one at home to help with wound care. The data exists, scattered across the PBM, the behavioral health carve-out, the hospital social worker’s notes, and the home health intake form. No federated network I’ve been part of synthesizes all of that in real time.

Why Previous Platforms Failed

The skepticism is earned. Let me engage with it directly.

Google Health launched in 2008 to let patients aggregate their own data. Shut down in 2012. Nobody used it because patients don’t want to manage health data. They want someone to manage their health.

Haven. Amazon, Berkshire Hathaway, and JPMorgan. Three of the most operationally capable companies in the world. Dissolved in 2021 with nothing to show for it.

The consistent pattern: technology overlaid on the existing system without changing incentives, relationships, or workflows. A portal nobody logs into. A dashboard glanced at weekly. An interoperability standard technically implemented and practically empty.

I’ll add one more failure mode that doesn’t get discussed enough: I watched a 14-site CIN deploy a care coordination platform designed to reduce readmissions. Six months in, readmissions hadn’t moved. Care manager burnout was worse. The platform was working exactly as designed — it surfaced thousands of care gaps nobody knew existed. Missed screenings, lapsed referrals, unfilled prescriptions, overdue follow-ups. The problem wasn’t the technology. The problem was that the organizational model couldn’t absorb the answers fast enough. Twelve care managers can’t act on 9,000 alerts. The platform found the problems. The humans were drowning in them.

That experience changed how I think about this. The technology layer isn’t just a detection system. It has to act, not just alert. And it has to act at a scale that human-only care management teams can never reach.

What Changed

Two things happened simultaneously.

AI can now engage patients directly in natural language, at scale, continuously. Not a portal. Not a decision tree. A conversational intelligence that can call a patient after discharge, understand they sound confused about medications, detect an unfilled prescription, notice a missed follow-up, and take action. Not just surface an alert for a care manager to handle on Tuesday. Act. This was not possible at a useful quality level before 2023.

Healthcare interoperability finally has plumbing that works. FHIR APIs are mandated. TEFCA is establishing national data exchange. CMS requires payers to share data programmatically. The data is still messy, still incomplete. But for the first time there are standard pipes instead of fax machines.

AI that can act on information, plus information that can be accessed programmatically. That combination creates the precondition for an intelligence layer that coordinates care across organizational boundaries without requiring those organizations to merge, affiliate, or even agree to coordinate.

The Honest Complication

I’d be dishonest if I framed this purely as an incentive problem, as if health systems and payers are strategically refusing to build the care layer because it threatens their boundaries. Some of that is real. An IDN that builds a layer coordinating care seamlessly across its own facilities and its competitors’ has made its organizational boundary less valuable.

But the more common reality is less dramatic and more frustrating: most health systems and payers are already trying to build this. They’re spending millions on care management platforms, population health tools, and patient engagement technology. They want coordination. They’re failing at it — not because they lack willingness, but because the technology until recently wasn’t capable enough, the interoperability wasn’t there, and the organizational complexity of deploying these tools across fragmented provider networks is genuinely, brutally hard.

The ceiling isn’t just strategic reluctance. It’s operational reality. And the platform that breaks through will have to be not just technologically superior, but operationally humble enough to work within the mess rather than pretending the mess doesn’t exist.

I’ve spent 42 posts describing how the healthcare system is organized. I haven’t yet described what needs to exist for the person the system was built to serve.

I’ve spent the last several months mapping how U.S. healthcare is actually organized. Thirty models across five layers, from the solo family physician to the $370 billion vertical empire of UnitedHealth Group.

The throughline I didn’t fully see until I was done: every organizational model in American healthcare was created to solve a coordination failure. And every one of them, in solving it, created a new boundary that produced a different coordination failure.

Physicians couldn’t negotiate alone, so they formed IPAs. The IPA solved the negotiating problem and created a data fragmentation problem. Hospitals and physicians needed unified managed-care contracts, so they built PHOs. The PHO solved the contracting problem but didn’t produce real clinical integration. Independent providers needed value-based contracts without merging, so they invented the CIN. The CIN solved the participation problem but created a governance problem.

Every solution works. Every solution leaves a gap. That gap creates the next model. It’s been running for forty years.

So where does this end? I see three possible futures. All three are already underway.

Future 1: Convergence Through Ownership

Optum employs or affiliates with roughly 90,000 physicians. CVS Health owns Aetna, Oak Street Health, Signify Health, MinuteClinic, and Caremark. Humana owns CenterWell clinics, home health, and pharmacy.

Let me be direct about this one: what these companies are doing is not integration. It’s annexation. They’re buying the pieces of the healthcare value chain one acquisition at a time and calling it coordination.

Inside the walls, it works. An Optum physician referring to an Optum specialist, with the patient insured by UnitedHealthcare, creates seamless coordination. Data flows. Incentives align.

But the math breaks at the edges. Even Optum, with 90,000 physicians, covers roughly 9% of the practicing workforce. The U.S. has over a million doctors. The solo family physician in rural Ohio isn’t getting acquired. The critical access hospital that’s the only facility for 50 miles isn’t joining Optum’s network. Vertical integration solves coordination for the patients inside the fortress and leaves everyone else with the same fragmentation they’ve always had.

And there’s a regulatory ceiling that’s getting lower every year. The FTC sued US Anesthesia Partners. The DOJ is blocking hospital mergers. State AGs are investigating PE-backed rollups. There is a political limit to how much healthcare a handful of corporations can own.

I respect the operational ambition. I don’t think it scales to a country of 330 million people.

Future 2: Federated Networks With Shared Risk

CINs, ACOs, and the enablement platforms that support independent physicians represent a more democratic alternative. Don’t merge. Align. Share data, follow common protocols, split the savings.

About 480 MSSP ACOs now cover 11 million Medicare beneficiaries. Clinically integrated networks let independent physicians participate in value-based contracts while keeping their practices. Enablement platforms provide the analytics and care management that small practices can’t build alone.

I have more sympathy for this future than the first one. It preserves physician independence. It doesn’t require massive capital. It works with existing structures.

But I’ve watched enough federated networks up close to know their failure mode. The physician champion who drove CIN participation retires and nobody replaces the energy. The health system anchor gets a new CEO who decides employed physician growth matters more than network maintenance. The shared savings pool has one bad year and three practices quietly disengage. I’ve seen a thriving CIN hollow out in 18 months after a single leadership change.

Federated networks are elegant when they work. They’re fragile by design, because they depend on sustained voluntary commitment from parties with different economics and different priorities.

Future 3: Platform-Mediated Coordination

This future doesn’t have a single company building it yet. It’s an architectural pattern emerging at the intersection of foundation models, ambient computing, and healthcare’s slow progress on interoperability.

The theory: if no entity can see the full patient picture across all 30 organizational models, the solution isn’t a bigger entity or a better network. It’s an intelligence layer that sits between the patient and the system. One that persists across care transitions, synthesizes data from multiple sources, and acts on behalf of the patient regardless of which organizational model they’re inside at any given moment.

Not an entity. Not a network. A layer.

It doesn’t require ownership, formal agreements, or physicians giving up independence. It works with the existing stack. And the technology to build it became viable at a useful quality level approximately 18 months ago.

The counterargument is earned: healthcare is littered with the corpses of technology platforms that were going to fix coordination. Google Health. Microsoft HealthVault. Haven. The pattern is consistent: technology overlaid on the system without changing incentives is a dashboard nobody uses.

This might be different. Or it might be Google Health with better language models. I’ll make the case for why I think it’s different, but I owe the graveyard its respect.

Where I Stand

I don’t think Future 1 gets us there. It concentrates power without reaching the majority of patients and providers who will never be inside a conglomerate’s walls.

I don’t think Future 2 gets us there alone. It’s the right instinct — align without owning — but the governance overhead and fragility are real constraints, not solvable objections.

Future 3 is the only architecture I see that scales without consolidation, works for independent providers, and puts the patient at the center. But it only works if it’s built on a payment model that rewards health over volume. The intelligence layer without value-based economics is an app. Value-based economics without the intelligence layer is a spreadsheet.

The endgame isn’t one of these futures. It’s the interaction between them.

But if the answer involves a platform, the immediate question is: why haven’t the entities that already exist just built one?

You’ve now read about 30 distinct organizational models across 9 categories. If your head is spinning, that’s normal. American healthcare’s organizational complexity is not a bug — it’s the natural result of a system that evolved incrementally over 100 years, with each new model created to solve a specific problem without replacing the ones that came before.

Here’s how to think about it all.

Layer 1: The Base — Where Physicians Practice

At the bottom are the practice models: solo practitioners, small groups, large single-specialty groups, multi-specialty groups, and faculty practice plans. This is where physicians actually see patients. Everything else is built on top of this layer.

Bottomline: About 50% of U.S. physicians still work in practices of 10 or fewer doctors. The base of American healthcare is smaller and more fragmented than most people realize.

Layer 2: The Connective Tissue — Networks and Integration

Above the practice layer sit the structures that connect independent providers into coordinated entities: CINs, IDNs, IPAs, PHOs, MSOs, and franchise/affiliation models. These exist because individual practices lack the scale, data, and infrastructure to participate in modern healthcare delivery and payment.

Bottomline: The spectrum from independence (IPA) to full integration (IDN) represents a fundamental tradeoff between autonomy and coordination. Most of American healthcare sits in the messy middle.

Layer 3: The Facilities — Where Care Happens

The physical infrastructure: AMCs, community hospitals, CAHs, safety-net hospitals, FQHCs, specialty hospitals, and physician-owned hospitals on the inpatient side. ASCs, urgent care clinics, and retail clinics on the outpatient side. SNFs, home health agencies, hospice organizations, and LTACHs on the post-acute side. CMHCs/CCBHCs, psychiatric facilities, and SUD treatment facilities for behavioral health.

Bottomline: Care is migrating from higher-cost to lower-cost settings — from hospital to ASC, from SNF to home health, from ER to urgent care. This migration is the dominant structural trend in healthcare delivery.

Layer 4: The Payment Models — Who Bears Risk

Layered on top of all the delivery infrastructure are the payment and risk structures: ACOs, ACO REACH, Medicare Advantage, Medicaid MCOs, risk-bearing provider organizations, and bundled payment programs. These don’t replace the delivery structures — they sit on top of them, reshaping incentives.

Bottomline: Payment models are the most powerful force reshaping healthcare organization. When you change how money flows, you change how care is delivered. Every organizational innovation in the past 15 years has been driven by the shift from fee-for-service to value-based care.

Layer 5: The New Entrants — Disrupting from Outside

Payer-provider convergence entities (Optum, CVS Health), DPC, concierge medicine, virtual-first primary care, and employer clinics represent models that route around the existing system’s limitations rather than working within them.

Bottomline: The most disruptive entrants are the vertical integrators — UnitedHealth/Optum, CVS/Aetna — who are assembling the full value chain under one corporate roof. They’re not playing within the existing structure. They’re building a parallel one.

The Government Parallel

Running alongside all of this are three massive government delivery systems — the VA, the Military Health System, and the Indian Health Service — that operate under completely different rules, funding mechanisms, and governance structures.

Why It Matters

The practical implication of all this complexity is that no single product, policy, or partnership strategy works across the entire healthcare system. A technology platform designed for an IDN won’t fit a solo practice. A payment model that works in California’s delegated market won’t work in a state where IPAs don’t exist. A care delivery innovation from an AMC won’t translate to a CAH.

Understanding which organizational model you’re dealing with — and which layer of the stack you’re operating in — is the prerequisite for doing anything effective in healthcare.

That’s the map. Use it well.

The physician still bills insurance for all clinical services. The retainer pays for the enhanced access and wellness services that insurance doesn’t cover.

What It Is

A concierge practice charges patients an annual retainer for premium access — but unlike DPC, it still bills insurance for clinical care. The retainer subsidizes small panels and extended services; insurance pays for the medicine.

Why It Exists

Affluent patients want guaranteed access, unhurried appointments, and a physician who knows them deeply. Physicians want smaller panels and better economics without giving up insurance revenue. Concierge medicine serves both.

The Tradeoffs

The upside: Exceptional access. Dual revenue streams (retainer + insurance). Reduced burnout for physicians. Patients love the experience.

The downside: Creates a two-tier system. When a physician converts to concierge, 1,500 patients lose their doctor so that 500 can have a premium experience. Retainer fees are not HSA-eligible. Limited to affluent markets.

The Bottom Line

Concierge medicine works for the people who can afford it. The ethical question — should access to your doctor depend on your ability to pay a retainer? — has no clean answer. But the model’s persistence and growth suggest there’s strong demand from both patients and physicians.

What It Is

About 1,800 DPC practices charge a flat monthly membership ($50–$150) for unlimited primary care. No insurance billing. No coding. No prior authorizations. Typically 400–800 patients per physician versus 2,000+ in traditional practice.

Why It Exists

Insurance-based primary care has become unsustainable for many physicians. Twenty-minute visit slots. 2,000+ patient panels. Hours of documentation for every hour of patient care. DPC strips all of that away and rebuilds the physician-patient relationship from scratch.

The Tradeoffs

The upside: Zero admin burden. Small panels. Longer visits. Transparent pricing. Physician autonomy. Evidence suggests DPC reduces total cost of care by keeping patients out of ERs and specialist offices.

The downside: Inaccessible to low-income patients. Doesn’t cover hospitals, specialists, or emergencies (patients need separate catastrophic coverage). Exists entirely outside the insurance system — no population health reporting, no quality measurement infrastructure. Inherently small-scale.

The Bottom Line

DPC is the purest expression of primary care idealism in American medicine. It proves that when you remove insurance overhead, primary care can work beautifully. The question is whether a model that serves 600 patients per doctor can address a country with 330 million people who need primary care.

Nuka exists because an Alaska Native tribal organization took control of its healthcare from the federal government and rebuilt it from scratch.

What It Is

The IHS is a federal agency providing healthcare to approximately 2.6 million American Indians and Alaska Natives through IHS-operated facilities, tribally operated programs (under Self-Determination contracts), and urban Indian health organizations.

Why It Exists

The federal government has a trust responsibility and treaty obligation to provide healthcare to federally recognized tribal nations. AI/AN populations face extreme health disparities — life expectancy 5.5 years lower than the national average — often in remote areas with no private healthcare infrastructure.

The Tradeoffs

The upside: Tribal self-determination empowers local control and cultural competence. The Nuka model is globally recognized. Community-based approach addresses social determinants in context.

The downside: Chronically and severely underfunded — IHS per capita spending is a fraction of other federal programs. Recruitment in remote locations is extremely difficult. Annual Congressional appropriations (not an entitlement) create funding uncertainty.

The Bottom Line

IHS is simultaneously the most underfunded federal healthcare system and the source of some of the most innovative delivery models in the world. The Nuka System of Care proves what’s possible when tribal communities have control, funding, and the freedom to redesign care from the patient outward.

These are not healthcare companies that offer insurance. They’re conglomerates that control the full value chain — from the premium dollar to the exam room to the pharmacy counter.

What It Is

A payer-provider convergence entity is an organization where a health insurer has acquired or built a provider delivery system, creating vertical integration across insurance and care delivery.

Why It Exists

When you control both the insurance premium and the care delivery, you eliminate the adversarial friction between payer and provider. You can direct patients to your own lower-cost sites of care. You combine claims data with clinical data. You capture margin at every step.

The Tradeoffs

The upside: Aligned incentives. Data advantage. Full value chain economics. Nearly impossible competitive moat.

The downside: When the company paying for care also delivers care, the conflict of interest is structural. Market power concerns are significant. FTC scrutiny is increasing. Organizational complexity is enormous.

The Bottom Line

Payer-provider convergence is the most powerful structural force reshaping American healthcare right now. If you work in healthcare and you’re not thinking about how Optum, CVS Health, and Humana are vertically integrating, you’re not seeing the board clearly.

What It Is

The VA is a direct-care, government-owned-and-operated delivery system. It’s the healthcare arm of the Department of Veterans Affairs, funded through Congressional appropriations, not insurance billing.

Why It Exists

The federal government has a legal and moral obligation to provide healthcare to military veterans. The private sector was never designed to serve the unique clinical needs of veterans: polytrauma, PTSD, traumatic brain injury, Agent Orange exposure, Gulf War illness, military sexual trauma, and veteran homelessness.

The Tradeoffs

The upside: Largest integrated delivery system in America. Unmatched expertise in veteran-specific conditions. Pioneered EHR and telehealth at national scale.

The downside: Access challenges (wait times, geography). Government bureaucracy. The Oracle Cerner EHR modernization has been plagued by delays, cost overruns, and patient safety concerns.

The Bottom Line

The VA is the proof of concept for integrated, government-run healthcare in America. Its successes (care coordination, research, population health) and failures (access delays, EHR modernization) offer lessons for everyone building large-scale delivery systems.

Under a bundled payment, the hospital receives one payment that covers the surgery and everything that happens in the 90 days after — including where the patient recovers and whether they end up back in the hospital.

What It Is

A CMS model where providers accept a single, prospectively set payment for an entire episode of care — typically a hospitalization plus 90 days of post-acute care.

Why It Exists

Fee-for-service reimburses each service independently with no coordination incentive. Bundled payments create accountability for the whole episode, particularly the post-acute period where the biggest cost variation exists.

The insight: two hospitals might charge similar amounts for a hip replacement surgery. But one discharges to home health (cost: $4,000) while the other discharges to a SNF (cost: $18,000). The bundled payment makes the hospital care about that difference.

The Tradeoffs

The upside: Creates accountability for the full episode. Incentivizes appropriate post-acute placement. Growing evidence of cost savings.

The downside: Providers bear risk for events outside their control. Requires sophisticated analytics. Post-acute partners resist utilization management that cuts their volume.

The Bottom Line

Bundled payments are the most intuitive value-based payment model: one price, one episode, one accountable entity. The biggest impact comes from post-acute optimization — getting patients to the right (lower-cost) recovery setting.

That’s a risk-bearing provider organization. It’s the endpoint of the value-based care spectrum.

What It Is

Any provider entity — medical group, IPA, health system — that accepts delegated financial risk from a payer. It receives capitation and manages total cost and quality for an assigned population.

Why It Exists

Payers delegate risk to providers because providers are closer to the point of care. When a medical group bears financial risk, it has direct incentive to eliminate waste, invest in prevention, coordinate care, and keep patients out of the ER and hospital.

The Tradeoffs

The upside: Maximum alignment of clinical and financial incentives. Provider controls utilization. Surplus from efficient management flows to the provider. Drives genuine care model innovation.

The downside: Financial risk — one bad flu season, one cluster of high-cost cancer cases, and the surplus evaporates. Requires actuarial, analytics, and care management sophistication. Creates incentive to undertreat.

The Bottom Line

Risk-bearing is not for the faint of heart or the under-capitalized. But it’s where the incentives are most aligned with keeping patients healthy. If you want to understand where healthcare is heading, study California’s delegated model — it’s been running this experiment for 30 years.

What It Is

A Medicaid MCO is a health plan that contracts with a state Medicaid agency to provide Medicaid benefits in exchange for a per-member, per-month capitated payment. Unlike MA (which CMS runs nationally), each state designs its own Medicaid managed care program.

Why It Exists

States adopted MCOs to make Medicaid budgets predictable (fixed monthly payments instead of open-ended fee-for-service), improve care coordination for medically complex populations, and leverage managed care tools for people with significant social determinant challenges.

The Tradeoffs

The upside: Budget predictability. Care coordination for complex populations. Can integrate physical health, behavioral health, and long-term services.

The downside: Capitation rates are often inadequate, squeezing provider reimbursement. Fifty different state programs mean fifty different rule sets. Prior authorization can impede access for vulnerable populations. Publicly traded MCOs extracting profit from Medicaid dollars is politically contentious.

The Bottom Line

Medicaid managed care is a $400+ billion market that doesn’t get the attention it deserves. If you’re building for healthcare’s most vulnerable populations, you need to understand how MCOs work, because they control the dollars and the networks.

What It Is

An MA organization is a health plan that receives a fixed monthly payment from CMS for each enrolled beneficiary and provides all Medicare Part A and B benefits. It bears full insurance risk. Many MA plans add supplemental benefits — dental, vision, hearing, gym memberships — to attract enrollment.

Why It Exists

Traditional Medicare is fee-for-service with no care coordination infrastructure. MA plans use managed care tools — provider networks, prior authorization, case management — to coordinate care and manage costs. The capitated payment model creates an incentive to keep people healthy.

When a health system owns its own MA plan (Kaiser, UPMC, Geisinger), it controls both the insurance premium and the delivery system. That’s maximum alignment.

The Tradeoffs

The upside: Full capitation incentivizes prevention. Supplemental benefits attract beneficiaries. Provider-owned MA creates total payer-provider alignment.

The downside: Prior authorization is a constant source of friction. Risk adjustment coding incentives create upcoding concerns. MA plans cost CMS more per beneficiary than traditional Medicare, which is politically contentious.

The Bottom Line

MA is the single most consequential trend in Medicare. Over half of beneficiaries have chosen it. It’s restructuring the economics of healthcare delivery for seniors. But the overpayment question — does CMS pay MA plans more than traditional Medicare would have cost? — is the political time bomb.

What It Is

A CMS Innovation Center model that allows organizations to take capitation-like risk for attributed Medicare fee-for-service beneficiaries. It includes explicit health equity requirements and is the successor to the controversial Direct Contracting model.

Why It Exists

MSSP’s shared savings were seen as too gentle to drive real transformation. REACH was designed to attract organizations willing to take full risk and bring capital, technology, and operational capabilities that traditional providers lack.

The Tradeoffs

The upside: Stronger incentives. Attracts investment. Health equity focus. Pathway to full-risk population health in traditional Medicare.

The downside: Non-provider participation raises concerns about profiteering from Medicare. Some entities have failed financially (CareMax went bankrupt). Critics call it backdoor Medicare privatization. Future administrations may kill the program.

The Bottom Line

REACH is where the policy tension between innovation and protection is most visible. It’s either the future of Medicare transformation or a dangerous experiment in privatization. The answer probably depends on who’s running CMS.

That’s it. That’s the ACO.

What It Is

An Accountable Care Organization is a group of providers that voluntarily takes shared accountability for the quality and total cost of care for a defined Medicare population. About 480 MSSP ACOs cover 11+ million Medicare beneficiaries.

The ACO itself is typically an LLC that sits alongside the participating provider organizations. It doesn’t own hospitals or practices — it coordinates contracts, quality reporting, and shared savings distribution.

Why It Exists

The ACA created ACOs in 2010 as the primary vehicle for moving Medicare from fee-for-service to value. The logic: if providers share in the savings from reducing unnecessary utilization and keeping patients healthy, they’ll invest in prevention and care coordination.

How It’s Organized

The ACO contracts with CMS and is responsible for a population of attributed beneficiaries. “Attribution” means Medicare looks at which primary care doctor a patient saw most often and assigns that patient to the doctor’s ACO.

Inside the ACO, the actual work happens through care managers, data analytics teams, and clinical protocols that try to prevent ER visits, reduce hospital readmissions, and manage chronic disease. The organizational infrastructure that does this work is often a CIN.

The Tradeoffs

The upside: Shifts incentives toward health. Encourages investment in prevention and care management. Can be built on existing structures.

The downside: Shared savings alone may not fund the transformation needed. Attribution is imperfect — patients can see any doctor they want, but the ACO is accountable for their costs. Administrative burden is significant.

The Bottom Line

ACOs are the most important structural experiment in Medicare. They’re not revolutionary — they’re evolutionary, layering shared accountability on top of fee-for-service. The results are modest but real: MSSP ACOs have generated billions in cumulative savings. The question is whether modest savings are enough to justify the complexity.

Walmart tried the same model at scale — and shut down all its health clinics in 2024.

What It Is

A retail clinic is a small healthcare facility inside a retail store (pharmacy, grocery) staffed by nurse practitioners or physician assistants. Limited menu: vaccinations, basic screenings, strep, UTIs, pink eye.

Why It Exists

A large portion of primary care visits are for simple, protocol-driven conditions that don’t require a physician. Co-locating with pharmacies maximizes convenience: diagnose and fill the prescription in one stop.

Retail clinics also serve a strategic purpose for their parent companies. CVS Health owns Aetna (insurance), MinuteClinic (primary care), and CVS Pharmacy (medications). The clinic is the entry point that connects a customer to the broader CVS Health ecosystem.

The Tradeoffs

The upside: Maximum convenience. Transparent pricing. Walk-in access. Integrated with pharmacy.

The downside: Very narrow scope. No ongoing patient relationship. Siloed from the patient’s medical records. Walmart Health’s closure demonstrated the model’s financial fragility when pushed beyond its niche.

The Bottom Line

Retail clinics work for simple, protocolized care. They don’t work as comprehensive primary care — Walmart proved that. The model’s future is probably as a strategic entry point for vertically integrated health companies (CVS/Aetna), not as a standalone business.

What It Is

Direct care at Military Treatment Facilities on bases, supplemented by the TRICARE insurance program that contracts with civilian provider networks. Operated by the Defense Health Agency.

Why It Exists

The military needs doctors who can deploy to combat zones, medics who can treat battlefield trauma, and hospitals that can surge during wartime. TRICARE supplements with civilian care where MTFs can’t meet all needs.

The Tradeoffs

The upside: Comprehensive coverage at minimal cost. Maintains medical readiness. Unified DHA improving standardization.

The downside: MTF quality varies. TRICARE network adequacy can be poor. Transition from military to VA care at separation is fragmented.

The Bottom Line

The MHS is healthcare designed for a purpose beyond healthcare: military readiness. That mission shapes its structure, its capabilities, and its limitations.

What It Is

Virtual-first primary care companies deliver most care through video, chat, and asynchronous messaging, with selective in-person access through owned clinics or partner locations. Amazon One Medical, Firefly Health, Carbon Health, and Galileo are leading examples.

Why It Exists

Traditional primary care has access problems: your PCP is booked three weeks out, only sees patients during business hours, and spends 15 minutes per visit. Virtual-first models use technology to provide faster, more continuous access.

The Tradeoffs

The upside: Dramatically faster access. Lower cost structure. Asynchronous messaging enables ongoing (not episodic) care. Attractive to younger, digitally native populations.

The downside: You can’t do a physical exam through a screen. Building trust and relationships is harder virtually. Most virtual-first companies aren’t yet profitable. State-by-state licensure requirements create regulatory complexity.

The Bottom Line

Virtual-first primary care proves that access is the biggest unmet need in primary care. The question isn’t whether digital care has a role — it clearly does. The question is whether it can replace the in-person relationship, or whether it’s a complement. The market is still figuring that out.