Every few years, a new technology category emerges in healthcare. It promises to change everything, to make everyone’s lives easier. And then it improves things at the margins—or worse, fixes nothing at all—without addressing the underlying problem.
I’ve been thinking a lot about why that keeps happening. And I think I finally have a clear answer.
Most technology vendors have been solving for the wrong thing.
Community-Based Care Has a Prevention Problem
Here’s the situation most community-based care organizations are in right now.
A clinician sees a patient. The session happens—a real, human interaction where important things are said, clinical signals are present, and the information needed to deliver great care, ensure compliance, and get paid all exist. Everything is right there, in the room, in the moment.
Then the session ends. And the work begins.
Notes get written—hours later. Compliance teams review them—days (or weeks!) later. Eligibility issues surface—weeks (or months) later, after the claim is denied. Clinical risk signals get missed—until a supervisor catches something in a chart review, sometimes several months after the visit. Two-thirds of improper Medicaid payments trace back to these failures that were preventable at the point of care.
The system doesn’t break in one place. It breaks across a chain of delays.
That’s the problem. Not that clinicians aren’t trying hard enough. It’s that the tools we’ve built are all designed to work after the moment of care—when the most important window to act has already closed.
Why Records and Intelligence Aren’t Enough
To understand where we’ve been stuck, it helps to name the categories.
For decades, healthcare ran on systems of record—EHRs, revenue cycle platforms, the infrastructure for storing what happened. Essential. Non-negotiable. But inherently backward-looking. By design, they capture care after it occurs.
Then came a wave of smarter tools—dashboards, analytics, ambient AI scribes. A real step forward. These systems of intelligence surface insights. They tell you what’s happening. But they stop there. Insight without execution still leaves the work to people. Someone has to read the dashboard, catch the compliance gap, and flag the eligibility issue. The burden is lighter, but it doesn’t disappear.
Both categories share the same limitation: They depend on what has already happened, and they still rely on humans to interpret and act.
The next evolution isn’t better records or better insights. It’s a system that acts.
Shifting Left to the Moment of Care
What we believe this industry is ready for—and what we are building—is a platform that shifts left, turning live clinical and operational signals into real-time action, within the flow of care. Not after it. Not in response to it. During it.
Here’s what that looks like in practice:
- A patient mentions a job loss during a session. Instead of that signal getting buried in a note, the system prompts the clinician to assess housing stability, screen for financial risk, and document the intervention—before the visit ends.
- A note is being completed that doesn’t meet compliance requirements. Instead of a CQI team catching it five days later, the system flags it in real time and guides the clinician to what’s missing—before sign-off.
- A coverage signal appears. Instead of a denial arriving three months later, the system helps uncover changes in eligibility verification—at the point of care.
The work doesn’t happen after the fact. It happens where it can actually change the outcome.
Eligibility verified before the visit. Documentation reviewed before sign-off. Coding validated against the actual encounter. Claims are certified before they reach the revenue cycle.
This is the shift from hindsight to foresight. From ambient scribing to agentic action. From a system that helps people do the work, to a system that does the work alongside clinicians and the teams that support them.
The Engine Behind the Action
I want to be specific about what makes this possible—because “the system that acts” is only useful if you understand how it all works.
The mechanism is interconnected AI-powered agents. Not autonomous tools making independent clinical decisions. Purpose-built agents that execute defined administrative workflows under your organization’s rules, consistently, at scale, without requiring a human to chase every step—while also not overriding the critical human-in-the-loop guardrails we’ve put in place.
Think of it as a layered system. Eleos’s core products—Documentation, Compliance, and our newly introduced RCM suite—capture and understand what’s happening across the care journey. On top of that sits an agent layer: Specialized agents for compliance, for revenue cycle, for documentation quality; each operating according to the policies and requirements your organization defines.
What makes this different from general-purpose AI is that these agents aren’t improvising. They’re executing. Your rules. Your requirements. Your clinical standards. Applied consistently across every clinician, every program, every site—without your team having to manually enforce them downstream.
That’s the architecture of a system of action. Not a tool that surfaces what needs to happen. A system that makes it happen.
What We Discovered Inside the Clinical Interaction
I want to share how we got here, because I think it matters.
We started Eleos by building what we believed was the best clinical documentation tool in behavioral health. And we did that. The outcomes validated it—peer-reviewed evidence showing greater reductions in depression and anxiety symptoms, clinicians saving hundreds of thousands of hours, and organizations holding onto staff they would have otherwise lost.
But as we went deeper into these clinical interactions, we kept seeing the same pattern. Even with better documentation, the downstream work remained. Compliance teams were still reviewing notes days and weeks later. Revenue cycle teams were still chasing denials months after care was delivered. Supervisors were still catching clinical risk signals in retrospective chart reviews.
The documentation improved. The delay didn’t.
That’s when something clicked: Because we operate inside the clinical interaction, we have something no downstream system ever could—full context, in the moment, before the window closes. We know what’s being said. We know what the organization requires. We can act on it before it becomes a problem.
That’s not a product feature. That’s a different model for what healthcare software can be. A system of action—one that doesn’t just record or surface what happened, but executes work in real time so that issues are prevented rather than discovered.
What This Means for the Organizations We Serve
For the behavioral health, SUD, care at home, and community-based care organizations we work with, this approach means a few concrete things.
Clinical teams get to focus on care. When the system handles administrative work within the flow of care—not piled on top of it—clinicians can be present with their patients. That’s what they signed up for. That’s what improves outcomes (and reduces no-shows.)
Compliance happens before it’s too late. Every note, reviewed in real time. Every gap, caught before sign-off. Not 5% of notes—100, uncovering millions of dollars in unidentified risk. Without scaling headcount.
Revenue is protected upstream. Eligibility verified before the visit. Documentation validated before the claim. Coding checked against the actual encounter. By the time a claim enters the revenue cycle, it’s already been certified. Organizations using this approach have prevented hundreds of denials and recovered six figures in previously missed revenue within months.
The system works for you—not the other way around. For too long, care organizations have served their software: Feeding it data, cleaning up what it missed, reacting to what it surfaced. A system of action flips that relationship. It executes on your behalf, under your organization’s rules, aligned to your goals.
A Different Kind of Platform
The signals needed to deliver better care, ensure compliance, and protect revenue already exist inside every clinical interaction. They always have. The question is whether your platform is capturing and acting on them in the moment—or piecing them back together after the fact.
That’s the distinction we’re building toward at Eleos.
Not a smarter scribe. Not a better dashboard. A system of action.
Where We’re Headed
We serve 270+ organizations across 39 states; these organizations trust us to help protect close to $5B in revenue. In 2025, our platform supported nearly 40,000 clinicians delivering care to over a million Americans.
We’ve saved clinicians the equivalent of 75 years of time—and that number means something to me, because time returned to a clinician is time returned to a patient. But time savings is where this starts, not where it ends.
The organizations we serve need more than efficiency. They need compliance that holds, revenue that doesn’t leak, and risk that gets caught before it becomes a crisis. That’s the promise of the system of action—and we’re only beginning to deliver on it.
That’s the system we’re building. And I believe it’s the one this field has been waiting for.
Ready to talk to an expert? Request a demo today. Want to geek out about our system of action? Shoot me a note.
