Most consequential journeys rarely announce themselves as such . While a truth I have seen reflected in some of the “deeper” things of life, on occasion the same reflections play out professionally as well.

I have been through a SOC 2 certification before and like most journeys taken in a hurry, the first time I was focused almost entirely on arrival. Policies were written. Evidence was collected. Controls were documented. The report was issued. We reached the destination on schedule and then, as organizations often do, we largely returned to where we had started.

It was only in reflection that I recognized what had been missed. Not the certification. We had acquired the certification. What was missing was everything the road might have offered had we been paying attention to it.

THE ROAD HAS A WAY OF REVEALING THINGS

There is a reason experienced travelers will tell you that the most valuable moments rarely happen at the landmark itself. They happen in transit, in the conversations that surface unexpectedly, in the detours that turn out to matter, in the discomfort that quietly reshapes how you see things.

Compliance programs work the same way .

SOC 2, at its surface, is a framework with defined Trust Service Criteria, audit requirements, and a report that either issues or does not. Those things are real and they matter. But the framework's most significant value is not in the destination it produces. It is in what the journey forces an organization to confront about itself.

Who actually owns this process? Does anyone know? Has this procedure ever been written down, or does it exist entirely in one person's institutional memory? What would happen to this control if that person left tomorrow?

These are not audit questions. They are organizational health questions. SOC 2 simply insists they be answered honestly.

WHAT I FOUND AT GoSB

When I joined GoSB and we began our SOC 2 journey, I arrived expecting the familiar pattern. What I found instead reframed how I think about these programs entirely.

The organization did not need to be persuaded to take the journey seriously. Team members wanted clearer ownership. Leaders wanted governance they could depend on as the company scaled. Departments wanted processes that lived in documentation rather than in memory. The appetite for that kind of rigor was already present. The organization simply needed a structured path to walk it out.

What struck me most was not any single milestone we reached. It was what surfaced along the way. Conversations about accountability that had never quite happened before. Decisions about ownership that had been deferred and were now being made. Processes being written down for the first time, not because an auditor required it, but because people understood, often for the first time, why it mattered.

Certification did not impose discipline on our team. It gave the organization's existing discipline a route to travel and a way to demonstrate when it had arrived.

THE MIRROR IN THE MIDDLE OF THE ROAD

SOC 2 is, in practice, a mirror. And mirrors are most useful not when they confirm what you hoped to see, but when they show you something you did not expect.

Organizations that approach the program as a checklist tend to see exactly that reflected back. Completed tasks, satisfied criteria, a report filed and forgotten. Organizations that approach it as a journey tend to discover something more durable. Gaps in accountability they did not know existed, governance habits worth building on, and operating foundations strong enough to support what comes next.

The difference is rarely about the framework. It is about the posture the organization brings to the road.

That posture shows up early and consistently. It shows up in whether the conversations center on what is required or what is right. It shows up in whether documentation is treated as an artifact produced for an auditor or as institutional knowledge built for the organization. It shows up in whether the program ends at the report or continues as a way of operating.

WHAT THE JOURNEY BUILT

We accomplished something genuinely impressive, not because we completed an audit, but because of what we built while pursuing it. 

Ownership became clearer. Decision-making became more defensible. Processes that had lived in people's heads became documented, reviewable, and improvable. The organization developed the habits that customers, partners, and investors expect from a company they are choosing to trust with their business.

That point is worth staying with. Trust is the underlying currency of SOC 2. The report is not the asset. The report is evidence of the asset. The asset is an organization that has demonstrated it can govern itself with consistency, transparency, and integrity, not once for an auditor, but as a matter of how it operates.

A NOTE FOR FELLOW TRAVELERS

As technology leaders, we invest heavily in systems, architecture, and tooling. Those investments are necessary. But scaling organizations eventually encounter constraints that no platform resolves and no infrastructure upgrade addresses.

Process breaks down. Governance lags behind growth. Accountability becomes diffuse precisely when it needs to be sharpest. Consistency erodes as complexity compounds.

Frameworks like SOC 2 are most valuable at exactly this moment, when speed feels more urgent than discipline, and when the temptation to bypass rigor in favor of momentum is strongest. The organizations that resist that temptation tend to emerge from the journey with something the destination alone could never provide. The operating habits required to scale with confidence and the institutional trust required to earn the right to do so.

I am proud of what we achieved. The certification is a meaningful milestone, and it signals clearly to the market that we take our obligations to customers and partners seriously. But what I will carry forward is not the report. It is the reminder that the most important things an organization learns about itself are rarely found at the finish line. They are found along the way, in the questions the journey insists you answer, and in the discipline it builds in the answering.

The destination confirms where you are, but the journey determines who you become.

Until the next dispatch, Dearest Gentle Readers, may your audits be ever in your favor.

‍

=================================================================================

‍

‍Larrian Martin is the Chief Information Officer at GoSB, a specialty healthcare RCM company. He was formerly Senior Vice President at Envision Healthcare and holds an Engineering Science degree from Penn State University where he studied in the Artificial Hearts Lab.

‍

_{WHAT WE CAN LEARN FROM FOUR RECENT STUMBLES}

‍

Late springtime of this year was a rough period for Team AI. The headlines arrived in a cluster, the way they always do when a hype cycle meets operational gravity. In the case the headlines revealed what can be fairly described, in the aggregate, as more in the realm of AI governance failures than technology breakdowns. 

‍

Starbucks scrapped an AI-powered inventory management system after the technology failed basic product identification — routinely confusing visually similar items like different milk varieties and missing stocked products altogether. A company promotional video from the launch period captured the problem plainly: a peppermint syrup bottle sitting on the shelf went unregistered as the system scanned the bottles on either side of it.

‍

 → Reuters — “Starbucks Scraps AI Inventory Tool” (via Yahoo Finance)

‍

A major Pizza Hut franchisee filed suit alleging that its AI-driven operations platform handed DoorDash delivery drivers unusual visibility into kitchen timing — allowing them to delay pickups, batch deliveries, and cherry-pick higher-tip orders. The result, per the complaint: pizzas sitting out longer, slower deliveries, and frustrated customers.

‍

 → Fortune — “Pizza Hut Franchisee Lawsuit: AI Adoption, DoorDash Delivery Drivers”

‍

Uber burned through its entire 2026 AI budget in four months after rolling out Claude Code to its engineering organization. By March, 84% of engineers were classified as agentic coding users. Roughly 70% of committed code originated from AI tools. About 11% of live backend updates were written by agents with no human in the loop. Monthly costs ran $150–$250 per engineer on average, with power users reaching $500–$2,000. The tools worked. The budget did not.

‍

 → Startup Fortune — “Uber Burned Its Entire 2026 AI Budget in Four Months”

‍

Amazon quietly shut down KiroRank — its internal leaderboard tracking usage of the company’s Kiro agentic AI platform — after discovering employees had learned to game it. The practice acquired its own vocabulary: tokenmaxxing. Engineers spun up unnecessary agents and ran pointless tasks to inflate their scores, burning real compute dollars in the process. Senior VP Dave Treadwell was eventually moved to state, explicitly: “Please don’t use AI just for the sake of using AI.”

‍

 → 404 Media — “Amazon Shuts Down Internal AI Leaderboard After Employees Cheated”

‍

If you have been waiting for the AI hype cycle to develop some visible cracks, this was your week.

‍

Read carefully, though. These are not the same story. They are four different failures — each with its own mechanism, its own accountability gap, and its own lesson.

‍

———

‍

STARBUCKS: THE VALIDATION PROBLEM

‍

Let’s be precise about what happened at Starbucks. This was not an adoption failure. It was not a change management failure. It was not a trust problem.

‍

The model did not work.

‍

A system deployed to manage inventory at one of the world’s largest retail operations could not reliably identify the products on the shelves it was scanning. A peppermint syrup bottle — sitting in plain sight — went unregistered while the system catalogued the bottles on either side of it. On milk inventory, performance was horrible in differentiating the distinctions of Oat, Soy, Almond, Skim, and Whole. The detail matters because it wasn’t a corner case. Reuters found errors were common. Imagine that, a company whose brand is tied to six versions of creamer rolls inventory AI that can't differentiate.

‍

The question that interests me is not why the model failed. Models fail. The question is how a model that failed this visibly cleared every checkpoint between vendor demo and production deployment at scale.

‍

That is a validation and oversight failure. Somewhere in the procurement and piloting process, the gap between “performs well on benchmark” and “works reliably on a Tuesday morning in a real store” was either not tested or not weighted appropriately. The technology broke. But something in the governance process let a broken technology get that far.

‍

In healthcare: We are observing similar gaps in prior authorization AI. Clinical decision support tools and AI-assisted prior authorization platforms which pass vendor demonstrations and procurement committees every day. The edge cases that break them — the ambiguous diagnosis codes, the dual-eligible patient, the non-standard formulary exception — rarely appear in the validation set. The peppermint syrup that went invisible on the Starbucks shelf becomes the authorization criterion the model confidently misclassifies. Providers and patients absorb the consequences. The procurement committee has moved on to the next initiative.

‍

 ———

‍

PIZZA HUT: THE SYSTEM BOUNDARY PROBLEM

‍

The Pizza Hut story is more structurally interesting than it first appears — because the technology worked exactly as designed.

‍

The platform gave delivery drivers visibility into kitchen operations and order timing. That visibility was presumably intentional — better information theoretically supports better coordination. What the designers appear not to have modeled is what a rational DoorDash driver would do with that information.

‍

A rational DoorDash driver, compensated per delivery and competing for higher-tip orders, will delay a pickup if batching it with another order improves their economics. They will cherry-pick. They will optimize for their own outcome, not the franchisee’s. This is not bad behavior. It is entirely predictable behavior from an actor whose incentives were never aligned with the system’s intended outcome.

‍

The failure was a system boundary failure. The designers mapped the organization. They did not map every actor the system touched. Gig economy workers sitting entirely outside the org chart were handed a tool that let them arbitrage kitchen operations in real time — and they used it.

‍

The lawsuit is the franchisee’s problem now. The lesson belongs to anyone deploying AI in an environment with multiple stakeholders who do not share the same incentives.

‍

In healthcare: Watch OpenAI’s stated ambition to own the patient advisor role. A platform that sits between patients and care navigation gains structural visibility into medication adherence signals, care-seeking behavior, and Rx decision-making before the provider or payer does. Prescription data is the most sensitive edge of that exposure. The regulatory frameworks for what a patient advisor platform can do with that visibility — and who is accountable when it is exploited — do not yet exist at the scale the technology is moving. The Pizza Hut franchisee did not anticipate what DoorDash drivers would do with kitchen timing data. The healthcare industry should be asking harder questions about what a patient-facing AI platform will do with longitudinal health intent data.

‍

 ———

‍

UBER: THE FINANCIAL GOVERNANCE PROBLEM

‍

The Uber story is, in some ways, the most uncomfortable of the four — because it is a success story with an unaccounted bill.

‍

The Claude Code rollout worked. Adoption climbed from 32% of engineers in February to 84% classified as agentic coding users by March. Roughly 70% of committed code originated from AI tools. About 11% of live backend updates were produced by agents operating with no human in the loop. By any engineering adoption metric, this was a successful deployment.

‍

And yet Uber burned through its entire annual AI budget in four months.

‍

The structural flaw was organizational, not technical. The teams driving adoption — engineering, product — were not the teams managing the spend. Monthly cost per engineer ranged from $150 to $250 on average. Power users ran $500 to $2,000. Naga, Uber’s engineering leader, reportedly spent $1,200 in a two-hour personal demo. That was not anomalous. That was the tool working as designed for the workloads it was built to handle: parallel agent execution, large-scale codebase refactoring, automated test generation.

‍

Nobody had modeled what successful adoption at scale would actually cost. And the leaderboard culture Uber built around Claude Code usage accelerated token burn directly.

‍

“The direct link to new consumer-facing features remains a mystery.”

 — Uber COO Andrew Macdonald

‍

Output went up. Outcomes remain unclear. That distinction — between measuring what the AI produced and measuring what the AI produced that mattered — is the load-bearing question that most AI business cases leave unanswered.

‍

In healthcare: Health system and RCM leaders are pushing AI adoption — ambient documentation, coding assistants, denial prediction tools — with urgency that is culturally driven as much as strategically driven. The engineering and operations teams measuring adoption velocity are rarely the same teams accountable for financial outcomes. Clinical ROI is at least as hard to attribute to AI output as Uber’s consumer-facing features were to token consumption. The CFO will eventually see the bill. The question is whether the governance structure to connect spend to outcome was built before or after that conversation.

‍

 ———

‍

AMAZON: THE INCENTIVE DESIGN PROBLEM

‍

Amazon’s KiroRank story is the oldest lesson in the group, wearing the newest clothes.

‍

KiroRank tracked how much employees used Kiro, Amazon’s agentic AI developer platform. The intent was presumably to encourage adoption, surface engaged users, and create visibility into organizational AI maturity. What it created instead was tokenmaxxing — a practice so widespread it required a named Senior VP to intervene publicly.

‍

Employees created unnecessary agents. They ran pointless tasks. They burned real compute dollars doing work that had no business value, optimizing for a metric that was supposed to measure business value. Dave Treadwell’s instruction — “Please don’t use AI just for the sake of using AI” — is the kind of sentence that only needs to be said when the measurement system has thoroughly decoupled behavior from intent.

‍

This is not a technology story. It is not even an AI story. It is a Goodhart’s Law story, playing out in a new medium: when a measure becomes a target, it ceases to be a good measure.

‍

The only thing surprising about it is that anyone was surprised.

‍

In healthcare: RCM organizations run on KPIs, and KPIs have always had this property. Optimize for DSO and watch teams prioritize easy receivables over complex recovery. Optimize for denial rate without tracking appeal recovery and watch the wrong denials get written off quietly. Optimize for AI tool utilization — logins, queries, recommendations accepted — and watch staff find the path of least resistance through the workflow without actually changing how they work. The downstream cost in healthcare is not compute spend. It is revenue leakage, compliance exposure, and the quiet erosion of outcomes that nobody’s dashboard is measuring.

‍

 ———

‍

FOUR FAILURES: FOUR DIFFERENT LESSONS

‍

It is tempting to collapse these into a single narrative: AI is overhyped, organizations are underprepared, management is failing. That narrative is not wrong. It is just not precise enough to be useful.

‍

The taxonomy matters:

‍

Starbucks was a validation failure. A model that could not perform its core function cleared procurement and reached production. The question to ask before the next deployment: what would it take to break this in production, and did we test for that?

‍

Pizza Hut was a system boundary failure. The designers modeled the organization. They did not model every actor the system touched. The question to ask: whose incentives does this system change, and are any of those people outside our org chart?

‍

Uber was a financial governance failure. A successful rollout produced an unaccounted bill because adoption accountability and cost accountability lived in different parts of the organization. The question to ask: who is responsible for connecting AI spend to AI outcomes, and do they have the authority to act on that?

‍

Amazon was an incentive design failure. The measurement system created the behavior it was intended to track — and that behavior was the wrong one. The question to ask: if someone wanted to game this metric, how would they do it, and is that what we are currently rewarding?

‍

The organizations that navigate this decade well will not necessarily be those with the most sophisticated models. They will be the ones that asked these questions before the go-live date rather than after the lawsuit, the budget overrun, or the SVP memo.

‍

 ———

‍

WHAT THIS MEANS IF YOUR SETTING IS HEALTHCARE

‍

Healthcare did not invent any of these failure modes. But it has a well-documented talent for making them more consequential.

‍

The validation gap that let a broken Starbucks model reach production is the same gap that allows AI-assisted prior authorization tools to confidently misclassify edge cases that experienced reviewers would catch. The procurement committee approved the accuracy rate. Nobody modeled the failure mode.

‍

The system boundary failure at Pizza Hut has a direct analog in any patient-facing AI platform with ambitions at the scale OpenAI , for example, is targeting. When a platform gains longitudinal visibility into patient health intent — including Rx behavior, care navigation, and adherence signals — before the provider or payer does, the question of who is accountable for how that visibility is used is not a future regulatory problem. It is a present governance problem.

‍

The Uber financial governance failure maps precisely onto AI rollouts where clinical and operational leaders are measured on adoption velocity and CFOs discover the cost structure in a quarterly review. Output is measurable. Outcomes, in healthcare as at Uber, are considerably harder to attribute.

‍

And the Amazon incentive design failure is native to RCM. DSO. Denial rates. Authorization turnaround. Clean claim rates. These metrics exist because they are measurable. What gets written off, delayed, or missed because the incentive pointed in a slightly wrong direction is considerably harder to see — and AI tools do not fix that problem. They accelerate it.

‍

The harder, slower, less glamorous work is the same across all four failure modes:

‍

1. Validate before you deploy. Not on benchmarks. On the edge cases that will actually break the system in production.

‍

2. Map every actor the system touches. Not just the ones on your org chart.

‍

3. Connect spend to outcomes before adoption runs ahead of accountability.

‍

4. Design incentives for the behavior you want, then ask how someone would game them.

‍

None of that is an AI capability. All of it is a leadership responsibility.

‍

===================================

‍

‍Larrian Martin is the Chief Information Officer at GoSB, a specialty healthcare RCM company, He was formerly Senior Vice President at Envision Healthcare and holds an Engineering Science degree from Penn State University, where he studied in the Artificial Hearts Lab.

‍

Last weekend, I stood at the finish line of the MDI Marathon in Bar Harbor and watched my sons complete a race they had trained for over many months. The preparation was grueling, the race itself was a struggle, and the finish brought both exhaustion and joy.

Two things struck me in that moment. First, these apples did not just fall far from the tree - they rolled down the hill, jumped the fence, and started growing in an entirely different orchard. Second, I realized that a marathon is the perfect metaphor for healthcare’s adoption of AI.

Training takes discipline, persistence, and more than a few blisters. The race itself demands stamina and adaptability. And crossing the finish line delivers the ultimate mix of exhaustion, relief, and proof that something hard, maybe even a little crazy, is absolutely possible.

WHY HEALTHCARE IS AHEAD OF THE CURVE

Healthcare is often labeled a “slow adopter” of technology. EHR rollouts took decades, interoperability has been a perennial struggle, and even simple digital tools have historically lagged. Yet in AI, something different is happening.

- A recent AMA survey found that two-thirds of physicians now report using AI-enabled tools, a jump of 78% since 2023.

- Menlo Ventures & Bain’s 2025 Healthcare AI Index shows 22% of healthcare organizations have deployed domain-specific AI. An amazing 7× increase in just one year.

- Forbes reported that 27% of health systems are leading with AI adoption, compared with only 9% of companies in other industries.

- In revenue cycle management, a Waystar study found providers using AI saw double-digit gains in cash acceleration and denial reduction.

Healthcare can’t afford to sit out the race. And what’s remarkable is that our industry, long criticized for being slow to adopt, is now out in front. That’s not an accident. It’s the result of endurance, grit, and a recognition that the stakes are simply too high to lag behind.

FIVE WAYS AI IS CREATING ADVANTAGE

1. Mapping the Patient Journey. AI connects fragmented data across the continuum — from the accident scene to emergency transport, from hospital care to recovery, and finally to billing and settlement. Seeing this journey as one story reduces delays, prevents denials, improves cash flow, and ultimately creates a smoother experience for patients and providers alike.

2. Driving Specialization and Consolidation. Not all claims behave the same. Motor vehicle accidents and workers’ comp claims are complex, document-heavy, and often contested. AI helps identify and route these cases at scale, enabling specialized management. Over time, this specialization leads to consolidation: fewer organizations handle them, but with sharper expertise and stronger results.

3. Boosting Clinician Productivity. Ambient documentation, predictive coding support, and decision-assist tools free up clinicians to spend more time with patients and less time buried in the EHR. AI helps reduce burnout while allowing medical staff to operate at the top of their license.

4. Reducing Administrative Costs. Revenue cycle, scheduling, utilization review, denial management, and prior authorization are loaded with manual work. AI-driven automation cuts out friction, lowers administrative overhead, and accelerates processes that historically consumed countless hours and dollars.

5. Improving Patient Outcomes. Beyond finance and operations, AI enables predictive analytics, early detection of risk, personalized care plans, and faster intervention.

By integrating insights across the continuum, providers can deliver more proactive, outcome-focused care.

LOOKING FORWARD

Like a marathon, adopting AI in healthcare is demanding and sometimes exhausting, but worth the effort. The discipline to train, the commitment to run, and the satisfaction of finishing strong are all part of the journey. And in this case, the finish line means proving that healthcare can move faster, adopt smarter, and lead with resilience. I’m proud of our industry for embracing this challenge, for showing the endurance to push through and the grit to prove that even a historically slow adopter can surge ahead when it matters most. We know we can do hard things. At GoSB, we’re bringing this AI-enabled approach directly to one of the most challenging areas in healthcare finance: accident-related claims.
‍

Larrian Martin is the Chief Information Officer at GoSB, a specialty revenue cycle management company, and formerly EVP of Data & Insights at Envision Healthcare.

To learn more about how specialization and automation can turn complexity into opportunity, email him at l.martin@go-sb.com.

‍

The decision to specialize in revenue cycle management is, at its core, a data-driven one informed by an understanding of scaling and innovation advantages. Trying to be good at everything is a recipe for underperformance — wisdom lies in knowing when to partner with a specialist to create success.

Accident-related claims — motor vehicle accident and workers’ compensation — represent less than 5% of overall volume for most physician providers and health systems, and less than 10% of overall volume for EMS agencies, yet they demand disproportionate resources and consistently underperform in recovery. Your healthcare billing group’s data will certainly confirm.

Generalist revenue cycle management (RCM) approaches to these claims underperform because they require unique data sets, workflows, and payer strategies. A specialist, however, can concentrate volume across many providers to achieve economies of scale that generalist shops simply can’t:

• AI and automation at scale — rules-driven steps and predictive models are more powerful when applied across millions of similar claims
  
• Aggregated data as an asset — denial patterns, payer behaviors, and recovery strategies become visible only when claims are pooled at scale
  ‍
• Systematic re-use of data — beginning with ambulance claims, a specialized partner can link the earliest touchpoint in the accident journey to downstream claims, creating early clarity and faster resolution

This is where GoSB’s leadership in the provider space and technical excellence shines. By dominating the patient journey entry point, we connect pre-hospital data with hospital based accident claim workflows — reducing duplication, accelerating payment, and building a more complete picture of patient care.

For executives, the takeaway is clear: accident-related claims are the classic example of the strategic decision to selectively add specialization partners. By handing over a small, high-value portion of claims to a focused ARC specialist, healthcare organizations capture a category with higher average payment rates and faster cashflow, creating a durable competitive edge over peers who default to a generalist approach.

In an industry where every decision carries risk, selectively partnering with a best-in-class ARC specialist is a no regrets strategy that positions providers for tomorrow.


Larrian Martin is the Chief Information Officer at GoSB, a specialty RCM firm, and formerly EVP of Data & Insights at Envision Healthcare.

‍

Hospital earnings calls starting on April 23rd are not going to surprise anyone paying attention to Washington DC or to the healthcare revenue cycle management ("RCM") vendors that have been signaling warnings for weeks.

The announcements are going to confirm what deliberate policy choices made in late 2024 and 2025 set into motion. The budget negotiations, the continuing resolution standoffs, and the government shutdowns revolved around healthcare policy. What opens now is the revenue cycle reading those decisions back to the industry, in the only language that registers at scale: the numbers.

WHAT WAS DECIDED, AND WHEN

It is worth being direct about the causal chain, because earnings commentary will not. 

The political drama of last fall conspired to produce a specific and predictable set of downstream consequences. Coverage grace periods tied to those negotiations began terminating on April 1, creating retroactive coverage loss on commercial claims. Medicaid disenrollment, which had been artificially suppressed through continuous enrollment protections, is now accelerating again. ACA participation, no longer propped up by the subsidy structures that stabilized enrollment through 2023 and 2024, is churning in ways the exchanges are only beginning to register. In short, a period of broad government policy-backed benefit expansion is in a reversal phase, gaining steam in its second full year.

None of this is mysterious. These were policy choices with known actuarial consequences, made by people who understood exactly what they were choosing. What is arriving now in hospital revenue cycles is not an emerging phenomenon. It is a scheduled outcome.

A SQUEEZE FROM BOTH SIDES

If payer mix erosion were the only spoiler, the story would be manageable. Health systems have navigated coverage churn before. But what makes this moment structurally different is that it is arriving simultaneously with a separate policy force moving in the same direction: site-neutral payment reform.

The CMS policy, now advancing with increasing regulatory confidence, establishes that the same procedure be reimbursed at the same rate regardless of whether it is performed in a hospital outpatient department or an independent physician office or ambulatory surgery center. For health systems that built their outpatient strategy around facility fee capture, this is not a marginal adjustment. It is a direct challenge to a core margin assumption and erodes years of strategic acquisition.

WHAT THE EARNINGS CALLS WITH ACTUALLY TELL US

The calendar lines up almost too cleanly. CHS and HCA report first, the week of April 23rd, offering the initial read on whether the signal is already measurable or still forming beneath the surface. Tenet and UHS follow a week later, where the story either confirms something undeniable or gets softened, explained, and rationalized into something more comfortable.

Watch the language as closely as the numbers.

Outpatient softness, if it appears, will be framed as transient. Emergency volume strength will be positioned as a positive. First quarter demand softness will receive more emphasis than it would in a normal first quarter, but systems are quietly building plans to absorb what they already know is coming.

Ensemble Health Partners flagged this dynamic explicitly in recent commentary, describing two forces beginning to collide: coverage quietly eroding while affordability, even for those who nominally remain insured, begins to give way. What is notable is that they are not describing a forecast. They are describing operational Q1 data already in motion.

In our last issue we touched on the topic of Claim Intensity. The payer response was a plan design shift to higher cost share which Ensemble is also noting as a headwind in first quarter 2026.

THE LAG IS THE RISK

The most important analytical point for health system leadership is one the earnings calls are structurally unable to communicate cleanly. Policy acts fast. Health systems absorb slowly.

The coverage terminations that began April 1 will take months to fully surface in volume and acuity data. The behavioral consequences of deferred diagnostics, avoided follow-ups, delayed elective procedures, will appear early as demand softness but accumulate quietly. The patients do not disappear. They defer. And when they return, they return sicker, under worse circumstances, and far less collectible.

What shows up in this earnings cycle is the leading edge of that curve. The patients who lost coverage in early January, plus those retroactively losing coverage in April. The procedures quietly falling off outpatient schedules in the past thirty days. The early movement in self-pay ratios that financial teams are watching but not yet ready to characterize definitively.

The full weight of the policy decisions made last fall has not yet landed.

THE DIFFERENTIATOR NO ONE IS TALKING ABOUT YET

The industry has a reliable habit of describing structural change as temporary until the moment it becomes undeniable. Payer mix erosion will be called a transition. Site-neutral headwinds will be characterized as regulatory uncertainty awaiting resolution. Demand softness will be attributed to seasonality, consumer sentiment, or post-COVID normalization, anything that preserves the narrative of a return to baseline.

But the baseline no longer exists in the same form. And as that reality firms up over the coming quarters, the divergence between health systems will not be explained primarily by clinical footprint, geographic market, or even payer contract strength.

It will be explained by the capability of their revenue cycle organizations to see what is happening, and help organizations reposition.

The RCM function is entering a highly strategic period.  Measurement speed and structural flexibility are genuine competitive variables. Organizations that navigate this environment most effectively are those that can detect payer mix shifts at the encounter level in near real time, reprice their financial assistance and presumptive eligibility logic dynamically, and redirect operational capacity toward the patient segments and service lines where margin is still recoverable. That is not a workflow optimization problem. It is an AI enablement solution -- usually requiring a specialized approach -- and the gap between organizations that have made that investment and those that have not is about to become visible in ways that quarterly filings will make difficult to ignore.

Equally significant is where the volume is actually going. The patient populations moving into self-pay, charity care, and Medicaid are not marginal to health system strategy. They are becoming central to it, both as a financial management challenge and as a mission obligation. The systems that treated these segments as administrative afterthoughts, managing them with legacy workflows, manual screening processes, and underfunded teams are going to find that the policy environment has made that posture untenable. These are no longer edge cases. They are the growth surface of an increasingly uninsured and underinsured market. Skills which find alternative third party revenue sources and pursue TPL cash will be differentiating.

The policy chapter may be closed but the financial chapter is just opening. The RCM organizations that can read it in real time while adapting faster than the earnings cycle moves are the ones that will define what health system financial resilience actually looks like on the other side of this.

====================================================================

Larrian Martin is the Chief Information Officer at GoSB, a specialty RCM firm, and formerly EVP of Data & Insights at Envision Healthcare.

‍

One might be forgiven for believing the great debate has already been settled. The discourse has been most animated, after all, on matters of volume and velocity. How many cases. How long they linger. Whether the machinery of adjudication can keep pace with the ambition of its design.

These are tidy questions. Respectable questions. But they are not the most interesting ones.

Because beneath the ledgers and timelines, a quieter narrative has begun to circulate. Not in formal reports, but in the corridors between them. In the follow ups. In the polite inquiries that arrive weeks after a decision has already been rendered.

Providers are winning, and still not getting paid. The payers are singing another anthem, united in writing another path forward.

Industry estimates now point to hundreds of millions of dollars in awarded claims that remain outstanding. Not under appeal. Not overturned. Simply… unfulfilled. Suspended somewhere between determination and delivery, as though the final act of the process were optional rather than implied.

Which invites a rather delicate question: If a system produces winners, but does not reliably produce payment, what precisely has been resolved?

The Independent Dispute Resolution process was introduced with admirable clarity of purpose. A structured forum. A neutral arbiter. A binding decision. A protected patient experience. A means, at last, of replacing protracted negotiation with something more orderly, more predictable, more fair.

And after three years of adjustments, these opening movements perform quite well. Cases are submitted. Evidence is considered. Awards are issued. Value is, in fact, determined.

But it is in the final movement, the one that follows the applause, where the composition begins to lose its discipline.

For once a determination is made, the system steps back. The burden of execution returns to the very parties whose incentives necessitated arbitration in the first place. And it is here, in this quiet transfer of responsibility, self-enriching behavior reasserts itself.

Payment timelines lengthen. Follow ups accumulate. What was meant to be a resolution begins, almost imperceptibly at first, to resemble a continuation, only now with more paperwork and less certainty.

And then, inevitably, one notices the asymmetry.

Providers operate within a regime of well-defined obligation. Even modest missteps can carry consequences, often in the range of ten thousand dollars per patient when enforced. The expectation is clear. Accuracy is not merely encouraged. It is required, and deviations are not without cost.

Payers, by contrast, are afforded a more flexible posture once an IDR determination is made. There is no equivalent, time bound enforcement mechanism that compels payment within a defined window. No automatic penalty that attaches to delay. No structural consequence that mirrors the discipline applied elsewhere in the system.

One side operates under defined penalties. The other, under optionality. And optionality, as it so often does, becomes strategic.

For even after an award has been issued, the matter does not always conclude. Claims may be reprocessed. Patient responsibility recalculated. Adjustments introduced that, while technically compliant, have the practical effect of reshaping the financial outcome.

The decision, one observes, remains intact. The economics, however, are subject to revision.

There is no need to assume ill intent to understand this pattern. It is sufficient to recognize the incentives at play. In a system where delay carries no immediate consequence, delay becomes a rational choice. Cash is retained. Timelines stretch. And a certain percentage of claims, through fatigue or pragmatism, are simply not pursued to their logical end.

Even the capital meant to support resolution does not always move with the urgency of the decision itself. Administrative fees and arbitration costs, held in what functions as escrow, are not consistently returned on timelines that reflect a system designed for clarifying finality.

In private conversations, concerns are beginning to surface about the handling of these funds at scale, with some suggesting that material sums may be sitting in limbo longer than intended. Whether operational backlog or something more structural, the effect is the same. Capital that should reinforce trust in the process instead introduces a second layer of uncertainty.

At a certain scale, delay begins to resemble something more than timing. And so the process completes, at least in theory. The outcome is known. The record reflects a resolution. Yet justice is not concluded.

Over time, such patterns have a way of shaping behavior. Smaller providers, already constrained, begin to question the value of engagement. Larger organizations grow selective in their pursuits. The theoretical promise of IDR begins to diverge, ever so slightly at first, from the cash that ultimately arrives.

And in that divergence, one finds the beginnings of a more structural concern.

For a system that determines value but does not ensure its delivery is not, in the strictest sense, resolving disputes. It is extending them under more formal terms, as an abandonment of the rule of law.

Even as policymakers, including physician legislators such as Greg Murphy, have devoted considerable effort to refining how IDR decisions are rendered, far less attention has been given to what becomes of those decisions once they are made.

It is perhaps in that omission that the system reveals its most telling characteristic. Adjudication has been defined with precision. Enforcement, rather less so. The rules describe how decisions are reached, but remain curiously understated on how those decisions are to be carried through.

In the absence of such structure, the market does what markets invariably do. It optimizes. Not for fairness. For advantage.

Which leaves regulators and policymakers with a question that is at once simple and rather consequential.

Is the purpose of this system to render decisions, or to ensure that those decisions carry respect? Because at present, the distinction is not merely academic. It is operational. And increasingly, it is financial.

The remedies themselves are not especially mysterious. Payment following an award could be automatic, governed by clear and enforceable timelines. Delays could carry consequences proportionate to their impact. Escrow processes could be bound by defined service expectations. Transparency could extend beyond the outcome of disputes to include adherence to those outcomes.

Such measures would not alter the spirit of the system. They would complete it. Until then, we remain in a rather peculiar arrangement. Where disputes are resolved on paper. But not always in practice.

And where victory, for many, is less a conclusion than an invitation to begin again yet another administrative task. One suspects the market will not remain so patient indefinitely.

‍

====================================================================

‍Larrian Martin is the Chief Information Officer at GoSB, a specialty RCM company, and formerly EVP of Data & Insights at Envision Healthcare.

‍

‍

‍

Innovation is part of the culture at GoSB. We are focused on realizing its full impact.
‍

For example, we framed a new approach to identifying auto accident claims -- also known as "accident-related claims" or "ARCs" in the industry -- based on the belief that the signals determining causation and financial responsibility are present, but not being assembled. As a result, qualifying claims pass through workflows undetected, and revenue leaks away.
‍

Over the past year, we have advanced that work. We test it at a scale only an ARC specialist can achieve and refine the architecture through real world deployment. What has emerged is a capability that consistently surfaces recovery opportunities that traditional workflows miss.
‍

Most revenue cycle systems rely on deterministic signals such as registration indicators, payer sequencing, or coverage discovery. These approaches are necessary, but they miss ARC cases because the signal rarely exists in a single place. Instead, it is often embedded in narrative and context, such as a paramedic report or clinical documentation that implies causation. Individually, these fragments may not change how an encounter is handled. Together, they provide clarity.
‍

We developed an ARC detection constellation that relies on emerging technologies and assembles distributed signals across claims data, clinical narratives, and coverage activity to identify cases that would otherwise remain invisible. In a provider evaluation across nearly 500,000 encounters, the model revealed meaningful recovery opportunities across the patient journey that existing workflows had never surfaced.
‍

What we are seeing now is just the first indication of what becomes possible when those signals can finally be assembled and understood.
‍

This is only the beginning.
‍

#######################################


Larrian Martin is the Chief Information Officer at GoSB, a specialty revenue cycle management company, and formerly EVP of Data & Insights at Envision Healthcare.

To learn more about how specialization and automation can turn complexity into opportunity, email him at l.martin@go-sb.com.

‍

‍

The term 'New Federalism' is being used to describe a new political philosophy that shifts power from the federal government back to the states. This is not a new concept but rather one that has grown out of the response to programs such as the New Deal and Great Society. We are beginning to see the changes put into effect, but we haven’t begun to experience the full import of these changes.

The enactment of the One Big Beautiful Bill Act (OBBBA) will result in significant changes to the Medicaid and Medicare reimbursement structure and funding. A potential change in the Affordable Care Act (as of this writing the extension of the ACA is still undecided) could increase the number of uninsured. A review of ground emergency medical transportation (GEMT) cost recovery and intergovernmental transfer (IGT) arrangements will be complete by the end of the year. If the provider tax does not survive scrutiny many states will no longer be able to continue GEMT or IGT funding, meaning these programs, essential to many EMS organizations across the United States, will close as well. This will close a crucial funding source for many agencies operating on the margin.

How we fund EMS systems in many communities is a delicate balance of the right volume of calls, coupled with a payor mix of Medicare and Medicaid (generally our largest group), third-party insurance patients, workers compensation, motor vehicle collision insurance, violent crime compensation, self-pay, and those who are uninsured and unable to pay. Any change in our payor mix can be a cause of joy or utter despair.

Claims denials for third-party insurers and Medicare Advantage are increasing. If changes in the ACA take place, coupled with a change in eligibility for Medicaid, many EMS systems will be challenged to field enough EMS units to handle requests for care. In some instances, they will close their doors.

EMS is not a business. We need to stop thinking of EMS as a revenue stream and realize that it is a cost center. EMS operates in a setting of inelastic urgent demand, absent consumer choice (when someone dials 9-1-1, we don’t provide a menu of EMS clinicians/providers), and skewed information, violating basic market preconditions. In every community EMS must be accessible regardless of ability to pay. This author has always said we cannot bill our way out of this problem. The societal value of EMS exceeds the private value, innovation markets fail absent public investment equally, and EMS' social returns dwarf what anyone could internalize, bill, and collect on. In EMS there are severe limits on consumer-driven or price-transparent models, and EMS faces these challenges head-on every day.

OPTIONS

Regardless of the outcome of the ACA and GEMT, with all the other changes that are occurring regarding insurers, EMS must strategize on how we continue to fund a system of care. The value of EMS lies in availability, readiness, and equitable access. Considering this—and the move to a New Federalism—EMS leaders, city and county administrators, and elected officials need to consider options in how to fund our EMS systems. Some areas worth exploration:

1. General Fund Allocations

This is the primary operating fund of a city, county, or state government. It pays for the basic functions of government that don’t have a dedicated funding source—public safety, parks and recreation, libraries, and community services. The money to fund these operating expenses typically comes from taxes, fees, and intergovernmental transfers. The challenge here is that everyone is trying to apportion their share of the same pot of revenue. Think about it like this: Do we buy new ambulances or renovate our playgrounds? That is the challenge faced by city administrators every day.

2. Enterprise Funds via Utility-Like Rate Structures

Local governments can treat EMS like a municipal utility, using revenues from other utilities (e.g. water or wastewater), setting up their own utility for EMS or setting user fees earmarked to cover EMS services diversifying beyond billing. Sun Prairie Fire and EMS in Wisconsin is a good example of this.

3. Special Districts

Creating EMS-specific or special-purpose districts funded through levies or assessments, often parcel-based, is another proven route. We see this with fire protection districts, but there are models of this in the EMS realm as well—Cambria Community Health District in California and Harris County Emergency Services District 8 in Texas are two excellent examples. The capability to set up a special district exists in every state. It provides a great degree of independence for EMS, but to quote Voltaire, “With great power comes great responsibility.”

4. Public Utility Model (PUM)

Here, the government contracts with private EMS providers to deliver services under strict performance-based contracts. The public retains control over assets and oversight, while operations are outsourced for efficiency and accountability. Richmond Ambulance Authority, EMSA in Oklahoma, and Sunstar in Florida are the leading examples of this concept.

5. Local Taxing Authority and Legislative Support

Certain jurisdictions are establishing broader taxing powers dedicated to EMS. For instance, Pennsylvania's Act 54 allows municipalities to increase EMS funding via local taxes, providing a more reliable revenue base.

RECOMMENDATIONS

We need to be responsible with the people’s money—not only our patients, but the public who may end up supporting us through some type of tax or district-based measure. This author has always said that you cannot have economic efficiency in EMS without technical/clinical efficiency. Take a deep dive into your organization and ask these questions: “Am I operating as efficiently and economically as possible?” “What are the full breadth of resources available to me?” “How can I do this better?”

1. Contracting and Cost-Sharing Arrangements

Municipalities can contract out EMS operations, share services across jurisdictions, or enter cost-sharing agreements to achieve economies of scale and reduce per-service costs. Contracting out vehicle maintenance to other services or creating purchasing agreements where multiple municipalities band to together for purchasing and better contract pricing for medical supplies, pharmaceuticals, and equipment, are all excellent places to start.

2. Billing

Billing will not go away for EMS; it will help defray the cost to the taxpayer for providing the service to the community. We have seen a rise in claims denial by private insurers and Medicare Advantage. Your billing contractor has probably spoken with you about improving documentation to improve claims processing and your accounts receivable. How many EMS organizations have spoken directly with the actual insurers? Go directly to Blue Cross. Ask them, “Please tell us what we need to do to be successful in this space.” Their answer may surprise you. If you submit claims for motor vehicles or workers compensation, these are specialties unto themselves. You probably aren’t collecting what you are entitled to recoup. This sub-specialty of EMS billing may require you to leverage an expert in this regard.

3. Low-Interest Loans, Grants, and State Programs

Many states provide funding support capital improvement grants, low-interest loans, matching grants, or training subsidies that can supplement local EMS budgets. Are you taking advantage of that currently?

4. Nonprofit Foundations/Philanthropic Backing

Seattle’s Medic One is a great example here, where communities can establish nonprofit foundations dedicated to EMS support, raising funds for training, equipment, and research through donations. There are others as well, and this type of support goes back to the founding of EMS. The Robert Wood Johnson Foundation in the 1970s funded EMS system development across the United States. The Gary Sinise Foundation and the Duke Endowment Foundation are two more fantastic examples of organizations currently helping EMS organizations.

5. Communicate Value

We need to communicate our value to the community. The police chief can talk about how his department has reduced crime. The fire chief can speak about how they have reduced fires in the community. Beyond response times, we must be able to clearly communicate to the public, healthcare stakeholders, elected officials, and government administrators how our efforts have reduced hospital length of stay and improved survival and functional outcomes for patients with major trauma, STEMI, CHF, asthma, COPD, pneumonia, and stroke. We need to demonstrate our value, and that we are more than just a ride to the hospital.

The coming shift toward New Federalism is more than a policy evolution, it’s a test of whether local leadership can safeguard EMS as a public good and an essential service in an era of fiscal contraction. Federal retrenchment through reduced Medicaid participation, provider-tax reform, or an altered Affordable Care Act will expose fragile funding systems already stretched by workforce shortages and escalating demand. But within that uncertainty lies the very opportunity John F. Kennedy described: to reimagine EMS financing through enterprise mechanisms, regional cost-sharing, dedicated districts, and public/private partnerships that value readiness, equity, and clinical capacity as essential infrastructure—not expendable expense lines.

EMS finance is, at its core, an exercise in moral accountability. Every dollar represents both a taxpayer’s trust and a patient’s lifeline. The challenge before EMS leaders is to navigate this changing fiscal landscape with innovation and integrity, building systems resilient enough to outlast political cycles and equitable enough to serve all who call for help. In doing so, we affirm that emergency medical services are not a business model to balance, but a covenant to uphold—one that ensures every community, regardless of wealth or geography, retains access to immediate, lifesaving care when it matters most.

‍

Daniel R. Gerard is the EMS Coordinator for Alameda City Fire Department and the immediate past president of the International Association of EMS Chiefs and a recognized authority in EMS system design and operations. With over 40 years of experience as a paramedic and registered nurse, Dan is a recognized EMS authority and has lead transformative EMS initiatives across the U.S. and abroad, including Hong Kong and the Bahamas. 

‍

What early HIPAA-ready moves reveal about enterprise healthcare AI

We are barely two weeks into the year, and healthcare AI has already crossed an important threshold.

Within days of each other, OpenAI and Anthropic released healthcare-focused announcements that move large language models out of experimentation and into HIPAA-aware, enterprise-ready territory. On the surface, the messaging sounds similar: HIPAA-ready AI, healthcare-grade security, and enterprise use cases. But just below that surface, the strategies diverge in ways that matter for health systems, RCM organizations, and healthcare technology leaders planning their 2026 roadmaps.
‍

CHATGPT AND CLAUDE SHARE A COMMON DEFINITION OF HIPAA-READY

HIPAA-ready does not mean “safe to paste PHI into any chat window”! At its core, HIPAA compliance in AI is not about the model. It is about the operating model around the model.

A HIPAA-ready AI offering requires the appropriate BAAs, Enterprise Agreements, data isolation and standard governance needed for roles, retention, and usage oversight.

In other words, AI itself does not become compliant, but we now have an available environment in which it can safely operate.  A governable industrial grade AI framework has finally arrived.

So, what will we do with this shiny new environment? From here the story radically shifts as Anthropic and OpenAI propose radically different solution paths and supporting ecosystems.  It can be a little bit overwhelming, so let's look at the big picture.
‍

A DIFFERENT WAY TO LOOK AT IT: FRONT-OF-HOUSE VS BACK-OFFICE GRAVITY CENTER

One helpful way to interpret the difference between these approaches is where each model naturally wants to live inside a healthcare organization. Their add-ons and announced acquisitions and integration partnerships tell the story.
‍
‍

CHATGPR: CONSUMER-FACING BY DESIGN, ECOSYSTEM-DRIVEN BY INTENT

ChatGPT’s healthcare strategy is not only clinician-adjacent, it's intentionally consumer-facing.

Beyond enterprise deployments, OpenAI is clearly investing in an ecosystem that supports patients actively bringing their own health data into the AI experience. Through a combination of recent acquisitions, enabling partnerships, and platform capabilities, ChatGPT is positioning itself as a trusted synthesis layer for consumer health information.

The direction is clear:

• Enable patients to share personal health data voluntarily
• Normalize AI as a place where individuals organize, interpret, and understand that data
• Deliver personalized insights that span conditions, history, and context rather than isolated data points

This is a meaningful strategic choice.

Rather than treating consumer health interaction as an edge case, ChatGPT appears to be embracing it as a core adoption vector; one that builds familiarity, trust, and daily usage long before a patient ever enters a clinical setting.

In that model, AI becomes:

• A personal health interpreter and coach
• A continuity layer across providers, encounters, and records
• A bridge between consumer understanding and clinical conversation

This reinforces why ChatGPT feels naturally suited to front-of-house healthcare experiences. Its strength is not just language, but synthesis—connecting disparate inputs into coherent, human-readable insight. When paired with governance and HIPAA-aware enterprise controls, that same capability scales into clinician workflows without losing its consumer roots.

The bet here is subtle but powerful:

If patients already trust an AI to understand their health, will clinicians and systems benefit from meeting them in that same shared context?

OpenAI is building the field of dreams with the understanding: If we build it, you will come. ChatGPT promotes a new business model for healthcare that doesn’t yet exist. An opportunity to enable and unlock informed consumer choice.
‍

CLAUDE: AGENTIC, WORKFLOW-FIRST, AND BACK-OFFICE ORIENTED

Claude’s healthcare positioning pulls in a different direction.

Rather than emphasizing a generalized enterprise assistant, Claude is framed around agentic, workflow-oriented automation, with specific healthcare use cases surfaced out of the box. Prior authorization support, coverage determination, benefits validation, and similar administrative workflows are central to the story.

This is an important distinction.

Claude is not just answering questions about healthcare processes. It is positioned to participate in those processes today

• Interpreting policy and coverage rules
• Reasoning through multi-step decisions
• Producing structured outputs that move work forward

Enabled with CMS integrations, and core data sources, Claude feel closer to an embedded operational agent than a conversational assistant. It aligns naturally with back-office domains as they exist today where decisions are repeatable, inputs and outputs are structured, and auditability matters as much as reasoning quality.

For organizations focused on revenue cycle operations, utilization management, or payer-facing workflows, this approach will feel immediately familiar. The assumption is that AI will live inside systems, not alongside them.
‍

WHY THIS DISTINCTION MATTERS

Healthcare organizations often struggle when they apply the same AI pattern everywhere.

Patient- and clinician-facing use cases benefit most from explainability, trust, and interaction. Back-office and administrative use cases benefit most from automation, consistency, and throughput.

Seen through this lens, the difference between ChatGPT and Claude is less about raw capability and more about where each fits best in the healthcare value chain.

The most mature healthcare AI strategies in 2026 will almost certainly use both patterns, and in some cases both tools, intentionally. What matters is identifying the initial partner best aligned with your core problem sets.

What the first 15 days of the year have made clear is this: healthcare AI has moved past the question of whether it can be compliant. The real debate now is where should the power of AI be focused.

That is a far more interesting conversation, and a strong signal of where the year’s great debate is headed.
‍

##################################################

‍

Larrian Martin is the Chief Information Officer at GoSB, a specialty revenue cycle management company, and formerly EVP of Data & Insights at Envision Healthcare.

To learn more about how specialization and automation can turn complexity into opportunity, email him at l.martin@go-sb.com.

‍

‍