What We Heard at Becker’s: The AI Conversation Is Growing Up
Summary: At the recent Becker’s Health IT + Digital Health + RCM Conference, the conversation around healthcare AI shifted away from mere adoption toward practical execution, change management, and tangible ROI rather than viewing technology as a total workforce replacement. Discussions included how healthcare providers are transitioning from basic retrospective dashboards to dynamic command centers that address root causes, particularly as payers deploy their own automated tools for post-payment recoupments. Ultimately, industry leaders emphasized that the true value of AI in revenue cycle management lies in combining specialized technology with human oversight to measurably improve cash flow and financial performance.
We spent a few days in Chicago at Becker’s Health IT + Digital Health + RCM Conference, and yes, there was a lot of talk about AI. But what was more interesting was how the conversation has changed.
There was less focus on Who has AI? and more on what happens after you have it. Much better questions are being asked around adoption, governance and return on investment (ROI).
Here’s what stuck with us.
Buying AI is the easy part
One of the simplest points we heard was also one of the best: implementation does not equal adoption. You can turn technology on but that doesn’t mean anyone is using it (or using it well).
A lot of the conversation centered on the human side of AI from change management to AI literacy to helping people understand how technology can make their jobs better rather than framing it as something that might replace them.
That resonates with us. The goal shouldn’t be to add AI everywhere you possibly can. It should be to take friction out of the work, and make better decisions faster. Automation isn’t meant to reduce headcount, but to shift back-office personnel from routine “doers” into strategic “thinkers” and problem solvers.
AI isn’t the goal. A better-performing revenue cycle is.
We may finally be outgrowing the dashboard
Dashboards are really good at telling you what already happened, rather than why issues occurred in the first place.
A theme we heard at Becker’s was the move toward something closer to a command center that explains why issues occurred and triggers immediate action or resolution. That’s a meaningful shift. Data becomes a lot more valuable when it helps someone act.
It’s also closely aligned with how we think about Auxo. Showing someone a problem is useful. Helping them understand the root cause, prioritize the work and determine the next best action is better.
The work is changing. People still matter.
For all the talk about automation, the more interesting workforce conversation wasn’t about eliminating people. It was about changing what we ask them to do.
If technology can handle more routine work, people can spend less time clicking, chasing and moving claims from one queue to another—and more time working exceptions, spotting patterns, solving problems and making decisions.
But there’s another side to that. If AI is doing real work, we need to manage it like real work.
We heard organizations talk about measuring AI quality and productivity with the same rigor they apply to people. That makes sense. Automation without accountability is still a problem. It just moves faster.
This is also where the move toward a more autonomous revenue cycle gets interesting. Autonomous doesn’t mean unattended. Technology can increasingly identify work, prioritize it and take action. Humans still provide oversight, judgment and intervention when something falls outside the norm.
Payers aren’t exactly standing still
Providers are getting better at automation. But so are payers.
One particularly interesting discussion was around the increase in post-payment recoupments. As providers improve front-end edits and denial management, payer activity doesn’t necessarily go away. Sometimes it simply moves further downstream.
We also heard examples of providers fighting automation with automation, including highly targeted appeals for things like ED downgrades.
And then there was this wrinkle: payers attempting to put language into contracts limiting providers’ use of AI and automation while continuing to automate their own processes. You can probably guess how providers feel about that. Welcome to the next phase of the battle of the bots.
But simply adding more automation isn’t the answer. Providers need to understand payer behavior, see patterns earlier and have the ability to respond quickly and intelligently.
Otherwise, we’re just finding more ways to play defense.
Be careful with anyone who says they do everything
Another point we appreciated was some healthy skepticism around the vendor that promises to do it all. RCM is complicated. That isn’t necessarily a flaw that needs to be engineered away. Putting every possible capability on a slide doesn’t automatically make someone an expert in all of them. Sometimes specialization matters.
We also heard some good discussion around outsourcing transparency. If you outsource work, you should know who is actually doing it. If your partner is simply turning around and outsourcing the work again, that matters.
More layers can mean less visibility, less control and more distance between you and the people responsible for moving your revenue. Whether you’re buying technology, services or both, the questions should be pretty straightforward:
- What problem are we solving?
- Who is accountable?
- And who is actually doing the work?
At some point, the conversation has to get back to cash
Revenue cycle professionals can talk about operational metrics all day. But your CFO probably doesn’t want to hear about work queues and success rates.
One message came through pretty clearly: translate the work into money. Cash flow, write-offs, true denial rates, revenue retraction rates, and the age of incoming cash (newer vs. older dates of service). Cash remains the single most critical metric.
The operational measures still matter. They tell you where to look and what to fix. But ultimately, the job of the revenue cycle is to collect revenue. Technology isn’t valuable because it automated 17 workflows or gave you a beautiful dashboard. It’s valuable if performance gets better.
Maybe the best sign is that the conversation is getting less exciting
We mean that in a good way. AI shouldn’t be the shiny new thing forever.
The conversations we heard in Chicago were increasingly about foundational elements: adoption, workflow, governance, accountability, payer behavior, workforce design and whether any of it is improving financial performance.
The next generation of revenue cycle isn’t going to come from choosing between people and technology. It’s going to come from figuring out how to use both exceptionally well.
We’re pretty comfortable with that future.