Healthcare operations still depend on people doing the work software was supposed to fix. While it works, it scales poorly. Every fragmented workflow requires more staff, training, and manual coordination.

Like most issues I write about, the problem lives in the messy middle, where we’ve automated the clean workflows, but the work between systems is still stubbornly manual (e.g., reading a PDF document, then checking a portal, followed by making a phone call, and then updating another internal system)..

In this article, I’ll break down why healthcare operations still scale like a 9-to-5 staffing business, how Champ AI is approaching AI-native workflows, and what this could mean for operations teams, physicians, and patients.

Healthcare has become more advanced clinically, but more complicated operationally.

We have better medications, more sophisticated diagnostics, more care sites, more insurance rules, and more technology touching every patient interaction. But every new layer creates more operational work, including medication approvals, eligibility checks, documentation, coordination, and handoffs.

Even as we’ve digitized parts of this work, the day-to-day still depends on people moving information between fragmented systems because the work is messy.

When one step breaks in this process, everything downstream slows. This frequently happens in clinic when a patient arrives without the right referral from their PCP, or when a medication is delayed because prior auth never received the supporting documents (happens all the time).

It’s administrative friction, but it’s also the connective tissue of healthcare.

The cleanest, ideal solution would be fewer fragmented systems, fewer payer-specific rules, better interoperability, and less administrative complexity in the first place. But that’s not the system we have… that’s only a fantasy. So healthcare organizations usually do the practical thing: they hire more people to keep the machinery moving.

The broader staffing trend speaks for itself. Healthcare has added a massive administrative layer around the actual delivery of care.

Some of that labor is necessary but a lot of it exists because healthcare still relies on people to bridge systems that do not communicate cleanly.

Hiring is slow and training is expensive. The work is repetitive enough to burn people out, but still messy enough that most automation tools struggle with it (too many edge cases). We’ve mostly automated the clean workflows that rely on structured data, clear decision paths, predictable steps, and minimal edge cases. Point solutions work well when the workflow is narrow and standardized.

But if a benefits check requires reading an intake PDF, logging into a payer portal, confirming details by phone, documenting the result, and writing it back into the EHR, the workflow becomes much harder to automate at scale.

This is the “messy middle” of healthcare operations. It’s the work between systems, after the clean API ends, where someone still has to read, click, call, verify, document, and escalate. But using people as middleware for this messy middle drives labor costs up, makes scaling harder, and makes standardization nearly impossible.

Champ AI is an AI operations platform trying to automate the messy middle of healthcare operations. Much of healthcare automation is still built around narrow point solutions. This may help with specific tasks, but it can also create more fragmentation when the underlying workflow crosses multiple systems. Champ AI’s core idea is that healthcare automation needs to be coordinated across multiple actions inside the same workflow not just one narrow step:

A simple example: an automation reads an intake PDF, logs into a payer portal, confirms benefits by phone, and writes the result back into the EHR or internal system.

This is the kind of workflow healthcare operators usually assign to trained staff because it crosses too many surfaces for traditional automation to handle cleanly.

The important part I want to emphasize is all of these actions—reading a pdf, putting the results into the EHR, making a phone call—can all happen under one streamlined workflow. Chaining the steps together in a way that can be reviewed, audited, and escalated when needed is where the operational value comes from.

While it’s reasonable to worry about errors in such a workflow delegated to AI, Champ AI has a control layer around the automation. There’s human-in-the-loop review, audit trails, self-healing browser agents, and deployment options like VPC or on-prem for organizations with stricter security requirements. This is key since healthcare operations have very little room for silent mistakes.

Eligibility verification is probably the cleanest example of how Champ AI works.

Here’s the current workflow:

Seven steps that could take 10–15 minutes. Multiply that by dozens of patients being scheduled per day, and then factor in the number of work days in a year… this becomes a huge staffing constraint.

The goal is moving humans from routine coordination to exception review–not removing them from the workflow entirely.

Champ AI’s pitch is that this kind of workflow should not require a human to manually coordinate every routine step. Their platform can read the document, navigate the portal, place the call when needed, and write the result back into the system, with humans pulled in for exceptions.

Champ AI is interesting because they’re going after the boring operational layer that determines how fast care can move forward. These are not the flashiest problems in healthcare. But as someone who experiences them as both a physician and a patient, I can assure you they shape the day-to-day experience of care.

If this kind of automation works, the impact is seen across operations teams, physicians, and patients.

For operational leaders (Directors/VPs of Ops), the value prop is right there: scale volume without adding human labor at the same pace. Emphasis is on repetitive work. This doesn’t mean removing operations teams but rather moving people away from repetitive coordination and toward exception management, quality oversight, and workflow redesign. The best operators in healthcare should not spend their days copying information between portals and internal systems. They should be focused on the edge cases that require human judgment.

For physicians like me, the impact is indirect but qualitatively better. There’s significantly less frustration with the administrative barriers when I’m ready to go with my clinical work. This means less prior authorization troubles that delay treatment, intake issues leaving gaps in the chart, and claims or coverage problems turning into patient frustration. The insurer may be behind a portal or phone tree, but the clinician is often the person sitting in front of the frustrated patient. Streamlined operations, as I’ve discussed repeatedly in Inefficiency Insights, remove avoidable friction that makes care harder than it needs to be.

For patients, the impact is much simpler. Fewer delays and less confusion. As a patient, I rarely know which operational step failed when I’m booking a new appointment or waiting for a medication. I just know the appointment was delayed, the medication was not approved, or someone told me to call back after I had already spent 40 minutes on hold. The goal of this kind of automation is to reduce that manual drag so patients experience care as faster, cleaner, and less fragmented.

Float Health and Pair Team are already taking advantage of Champ AI’s tools. Float Health, for example, is using Champ AI’s voice automation, while Pair Team is using Champ AI’s reliable browser automation for eligibility check. These are exactly the kind of portal-heavy workflow that tends to break traditional automations. 

As Champ AI scales and the market shifts to end-to-end operational workflows rather than point solutions, here’s what I’ll be watching:

In summary, Champ AI is going after one of healthcare’s least glamorous but most important bottlenecks: the manual operations work that sits between portals, PDFs, phone calls, and internal systems. If healthcare organizations want to grow without simply adding more people to move information around, they need automation that can handle the messy middle of the workflow. Champ AI’s focus is AI that’s capable of doing that work with enough flexibility, oversight, and auditability to matter.

If your team is trying to grow volume without adding headcount at the same pace, you can learn more or book a demo at here.

About this post: This is a partnered deep dive with Champ AI. The framework and conclusions, however, are my own. As a reminder, I only partner with companies working on problems I’d write about anyway, and Champ AI focus on automating the messy operational work between portals, PDFs, phone calls, and internal systems fits that bar. As always, my goal is to give you a transparent breakdown of what works, what still needs scrutiny, and why it matters for patients, physicians, and the health system. If you or your company would be interested in a partnership like this, click here.