The early hurdles were real and honestly pretty predictable in hindsight, but that doesn't make them any easier to navigate in the moment.

The first was buy in. Before you can build anything, you have to convince people that the investment is worth it. Early on, the ROI potential of automation isn't always obvious, especially when you're still proving the model. Getting leadership aligned around the value before you have a track record to point to requires a lot of credibility and even more patience.

Team capability was the other early challenge. Automation programs don't run themselves and we had to be honest about what our team could realistically build, support, and maintain. That assessment shaped a lot of early decisions around resourcing and priorities.

But the hurdle I underestimated most was pipeline. A sustainable program needs more than a handful of good ideas floating around. You need a real methodology for capturing opportunities, vetting them against clear criteria, and making sure only the highest value work gets prioritized

and built. Without that structure, you end up chasing whatever is loudest in the room rather than what actually moves the needle. Getting that intake and prioritization process right early was one of the most important things we did to set the program up for long term success.

Early on the answer was pretty straightforward. Processes with largely deterministic workflows were the best candidates for traditional RPA. A good portion of work in the PFS space fit that profile, as did a lot of our denials work. Repetitive tasks, minimal decision support, clear rules based outcomes. Those were the sweet spots and we had plenty of them to work through.

But as our skills grew and our technology expanded, we found that the boundaries of what automation could handle kept pushing further than we expected. Work we initially assumed required human judgment turned out to be highly automatable once we took the time to properly map and standardize the process. The ceiling kept moving and it still is.

The areas where we consistently found that a human touch remained essential came down to a few things. When a task requires someone to truly think through a complex or nuanced situation, when a patient experience is at risk, or when communication and empathy are critical to a quality outcome, those are the places automation has not been able to fully replicate. At least not yet. The technology is evolving fast and I'm careful not to close any doors permanently, but for now those are the spaces where we keep humans at the center intentionally.

Navigating this required intentional thought from the start, and a lot of it was anchored by a guiding principle the organizations that I have led in this space established early. We made a deliberate decision to put our EHR at the forefront of every automation decision. Leverage what you have, use it to its fullest capability, and protect the seamless experience it provides. That principle shaped everything that followed.

Before any automation was built, we focused on the workflow first. Redesigning and standardizing the process ahead of implementation was non-negotiable. You can't integrate automation into a workflow that isn't clean and that work required honest

conversations across operational and IT teams about how things were actually running versus how they should be running.

That cross functional engagement wasn't optional, it was baked into how we worked. Every automation considered outside of Epic had to be evaluated against existing workflows and systems to make sure it fit cleanly into the broader operational structure. That scrutiny slowed things down at times, but it also saved us from the bolt-on solutions and disconnected workflows that tend to derail programs down the line.

Building automation that truly fits your operational structure takes more upfront work. It also tends to last.

The biggest measurable impacts have come from productivity gains, on the front end with auth and referral statusing and on the back end with claim and denial work. Organizations that have leaned into statusing automation have captured significant gains, and those that have taken it further with customized solutions have seen it make an even greater impact to expedited cash collections.

When Automation is done well it also undoubtably reduces the need for net new headcount as volumes grow and that's a cost avoidance story that both finance and revenue cycle leaders can appreciate. Every organization will measure the value of automation differently. But cost savings is ultimately a byproduct of doing automation right.

The two things I keep coming back to are leadership buy-in and measuring the right things, and honestly, you can't have a sustainable strategy without both.

On the leadership side, you need more than just approval to get started. You need leaders who are genuinely invested in the outcome, people who will actively help remove the barriers that come up as you scale and refine your approach. Because they will come up. Automation isn't a build it and forget it thing, and if leadership isn't engaged, the first sign of friction can stall the whole effort.

The other piece is measurement, and I'd take it a step further than just tracking your own metrics. You need all your stakeholders aligned on what success looks like before you build, not after. Because everyone has a different definition of "success" and if you're not measuring the same things, you'll never fully agree on whether something is delivering value. There's that saying, you can't improve what you can't measure, and in automation it's especially true. Every strategy is ultimately trying to move the needle somewhere in the revenue cycle, and if you can't show that it's doing that, it's hard to justify scaling it.

Get the right people behind it and agree on how you'll measure it. Everything else gets a lot easier from there.